Citation record for Incident 72

Suggested citation format

Anonymous. (2017-10-17) Incident Number 72. in McGregor, S. (ed.) Artificial Intelligence Incident Database. Partnership on AI. Retrieved on November 27, 2021 from incidentdatabase.ai/cite/72.

Incident Stats

Incident ID
Report Count
Incident Date
72
27
2017-10-17

Tools

All IncidentsDiscover

Incidents Reports

Facebook auto-translation AI error results in Palestinian man's arrest

privacyinternational.org · 2017

A mistake in Facebook's machine translation service led to the arrest and questioning of a Palestinian man by Israeli police. The man, a construction worker on the West Bank, posted a picture of himself leaning against a bulldozer like those that have been used in hit-and-run terrorist attacks, with a caption that correctly translates to "good morning". Facebook's AI translated it into "hurt them" in English or "attack them" in Hebrew. Based on that, police officers arrested him later that day and questioned him for several hours before realising their mistake. The large number of dialects in use around the world means that Arabic is particularly difficult for machine translation services to handle, and mistakes are a regular occurrence. Facebook has apologised for the error.

https://www.haaretz.com/israel-news/palestinian-arrested-over-mistranslated-good-morning-facebook-post-1.5459427...

Facebook auto-translation AI error results in Palestinian man's arrest
Israeli police arrest Palestinian over Facebook post

dailymail.co.uk · 2017

Israeli police mistakenly arrested a Palestinian who posted 'good morning' in Arabic online which Facebook wrongly translated as 'attack them'.

The man uploaded a picture of himself leaning against a bulldozer at the Israeli settlement of Beitar Ilit, where he works, in the occupied West Bank.

In the caption, he wrote an Arabic term meaning 'good morning', but a software malfunction translated it to mean 'attack them' in Hebrew and 'hurt them' in English.

The man uploaded a picture of himself leaning against a bulldozer at the Israeli settlement of Beitar Ilit, where he works, in the occupied West Bank

In the caption, he wrote an Arabic term meaning 'good morning', but a software malfunction translated it to mean 'attack them' in Hebrew and 'hurt them' in English

Police only confirmed that a Palestinian had been mistakenly arrested then released following suspicions of incitement, but a report in Haaretz newspaper provided further details.

Officers believed the man was going to carry out an attack, so swooped in to arrest him.

It was unclear how such a translation error could have been made as there are no apparent similarities between the Arabic expression used for 'good morning' and the phrases in Hebrew or English.

Police were notified and the man was arrested last week, the report said.

He was released after a few hours when police realised the mistake.

Police spokeswoman Luba Samri said: 'A few days ago, a Palestinian was detained for questioning on suspicion of incitement through his Facebook page.'

She said he was 'immediately released' after the suspicions turned out to be false and the Facebook post has since been deleted....

Israeli police arrest Palestinian over Facebook post
Israel Arrests Palestinian Because Facebook Translated 'Good Morning' to 'Attack Them'

haaretz.com · 2017

The Israel Police mistakenly arrested a Palestinian worker last week because they relied on automatic translation software to translate a post he wrote on his Facebook page. The Palestinian was arrested after writing “good morning,” which was misinterpreted; no Arabic-speaking police officer read the post before the man’s arrest.

The Facebook post that mistranslated 'good morning' to 'hurt them'

Last week, the man posted on his Facebook page a picture from the construction site where he works in the West Bank settlement of Beitar Ilit near Jerusalem. In the picture he is leaning against a bulldozer alongside the caption: “Good morning” in Arabic.

>> Revealed: How Israel jails Palestinians because they fit the 'terrorist profile' >>

The automatic translation service offered by Facebook uses its own proprietary algorithms. It translated “good morning” as “attack them” in Hebrew and “hurt them” in English.

Arabic speakers explained that English transliteration used by Facebook is not an actual word in Arabic but could look like the verb “to hurt” – even though any Arabic speaker could clearly see the transliteration did not match the translation.

But because of the mistaken translation the Judea and Samaria District police were notified of the post. The police officers were suspicious because the translation accompanied a picture of the man alongside the bulldozer, a vehicle that has been used in the past in hit-and-run terrorist attacks. They suspected he was threatening to carry out such an attack and the police arrested him. After he was questioned, the police realized their mistake and released the man after a few hours.

The Judea and Samaria District police confirmed the details and said a mistake in translation was made, which led to the mistaken arrest. The police agreed the correct translation was “good morning.”

Keep updated: Sign up to our newsletter Email * Please enter a valid email address Sign up Please wait… Thank you for signing up. We've got more newsletters we think you'll find interesting. Click here Oops. Something went wrong. Please try again later. Try again Thank you, The email address you have provided is already registered. Close

The Palestinian man declined to speak with Haaretz. He removed the post from his Facebook page after the arrest....

Israel Arrests Palestinian Because Facebook Translated 'Good Morning' to 'Attack Them'
Palestinian arrested by Israeli police after Facebook mistranslated his 'good morning' post

ibtimes.co.uk · 2017

A Palestinian man in the West Bank was arrested by Israeli police after he wrote "good morning" in Arabic in a Facebook post that was mistranslated by the company's automatic translation software. Construction worker Halawim Halawi had reportedly posted a photo of himself holding a cup of coffee and a cigarette and smiling next to a bulldozer in Beitar Ilit, an Israeli settlement where he works, along with the caption.

Facebook's automatic translation software, however, interpreted the "good morning" post to mean "attack them" in Hebrew and "hurt them" in English, Israeli news site Haaretz reported.

Israeli police became suspicious of the post since he was standing next to a bulldozer, a vehicle that has been used in earlier terror attacks, the website reported.

Halawi was arrested under suspicion of incitement and questioned by police. He was released hours later after the mistake was realised. No Arabic-speaking police officer was consulted before the arrest was made, local media reported.

According to The Times of Israel, there is only difference in lettering between the colloquial Arabic phrase for "good morning to you all" and "hurt them."

The Facebook post in question has since been deleted.

The error comes after Facebook announced in August that it shifted to neural machine translation, which uses convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to automatically translate content across its site. Many tech companies including Facebook, Google and Microsoft have been pivoting towards neural machine translation and away from phrase-based, pattern-tracking statistical machine translation (SMT) to quicken and improve their translation software.

Still, automatic translation software been far from perfect and has resulted in numerous errors in the past.

#Israel arrested worker on occupied West Bank because Facebook translated 'good morning' to 'attack them' https://t.co/zqePwJBLn5 #Palestine — Raimo Kangasniemi (@rk70534) October 22, 2017

In August last year, Microsoft came under fire with many social media users calling for a boycott of the company's products after its Bing search engine mistakenly translated Daesh, the Arabic acronym for the Islamic State (Isis) into "Saudi Arabia."

In January 2016, Google Translate accidentally translated to Russian the Ukrainian word for "Russian Federation" as "Mordor" – the name of the evil, fictional region in JRR Tolkien's Lord of the Rings. The word "Russians" was also translated to "occupiers."

At the time, Google said its software looks for patterns in hundreds of millions of documents to decide and generate the best translation, but noted that the process is still difficult since the meaning of words depends the context in which they are used....

Palestinian arrested by Israeli police after Facebook mistranslated his 'good morning' post
Man arrested because of a Facebook mistranslation

alphr.com · 2017

A Palestinian man was arrested by Israeli police last week, after a post he left on Facebook saying "good morning" in Arabic was taken to meaning "attack them" in Hebrew.

A spokeswoman for the Israel Police’s West Bank district confirmed to The Times of Israel on Sunday that the man, a construction worker living in the West Bank, was arrested “on suspicion of incitement”. The man was held in custody briefly, but was released as soon as the mistake had been realised.

There is one letter’s difference between the Arabic phrase for "good morning to you all" and the Hebrew for "attack them", the report says. Facebook had automatically translated the post from Arabic, and no Arabic-speaking officer read it before making the arrest.

The post also contained a picture of the man holding a cigarette standing next to a JCB. The authorities took this to be threatening, when combined with the mistranslation, because bulldozers have been used in the past in terror attacks. The post has since been deleted.

Facebook has recently changed the way it translates between languages, but it seems the new method is not error-free. The company announced in August it had shifted towards neural machine translation. This uses convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to automatically translate content, instead of phrase-based statistical translation. This was expected to quicken and improve the translations.

"Our problem is different than that of most of the standard places, mostly because of the type of language we see at Facebook," Necip Fazil Ayan, engineering manager in Facebook’s language technologies group, said at the time. "We see a lot of informal language and slang acronyms. The style of language is very different."

But this is not the first time an online translation tool has caused problems. Last year, Google translated the Russian Federation to 'Mordor', the fictional land in Lord of the Rings using its automatic translation software. Because language depends so much on context, Google said, mistakes like this were bound to happen occasionally....

Man arrested because of a Facebook mistranslation
Palestinian arrested for saying ‘good morning’ on Facebook

postcourier.com.pg · 2017

Following a goof up by Facebook, a Palestinian man was arrested by the Israeli police for posting ‘Good Morning’ on the social media site.

According to reports, the man, a construction worker said “Good Morning” in Arabic in Facebook post, but the post was mistranslated to read “attack them” in Hebrew.

Following the mistranslation, the police said that they held the man under suspicion of incitement briefly, but that the man was released as soon as the mistake was realised.

The Palestinian construction worker was identified as Halawim Halawi and he is said to have posted a photo showing himself standing next to a bulldozer at the Israeli settlement of Beitar Ilit, where he works, in the occupied West Bank.

Previously, local reports noted, such vehicles have been used in terror attacks and officers thought he may be threatening to carry out such an atrocity.

Reports also pointed out that there is only one difference in lettering between the colloquial Arabic phrase for “good morning to you all” and “hurt them.”

Further, officers wholly dependent on Facebook’s automatic translation and did not consult an Arabic-speaking officer before making the arrest.

Law enforcement problems due to automated translation have affected tech companies and their popularity in the past.

Last year, Google’s online translation tool rendered the Russian Federation as ‘Mordor.’

Morder is the name of a fictional dark and evil region in JRR Tolkien’s Lord of the Rings books.

Further, “Russians” was translated to “occupiers” – causing more shame for the company.

At the time, Google argued that Translate worked by looking for patterns in hundreds of millions of documents – but translation remained difficult as the meaning of words was tied to the context in which they were used....

Palestinian arrested for saying ‘good morning’ on Facebook
Israel police mistakenly arrest Palestinian man for writing ‘good morning’ on Facebook

independent.co.uk · 2017

A Palestinian man who posted “good morning” in Arabic on Facebook was arrested after Israeli police mistakenly translated his message as “attack them”.

The man, who has not been named, posted a picture of himself working on a construction site in the West Bank settlement of Beitar Ilit near Jerusalem.

Alongside the photo, which shows the worker leaning against a bulldozer, he wrote: “Good morning.”

We’ll tell you what’s true. You can form your own view. From 15p €0.18 $0.18 USD 0.27 a day, more exclusives, analysis and extras.

Yet the automatic translation service offered by Facebook, which uses its own proprietary algorithms, translated the words to mean “attack them” in Hebrew and “hurt them” in English.

The police were notified and arrested the man on suspicion of incitement, but he was released after a few hours once the mistake was realised.

No Arabic-speaking officer was consulted before the arrest was made.

Shape Created with Sketch. World news in pictures Show all 50 left Created with Sketch. right Created with Sketch. Shape Created with Sketch. World news in pictures 1/50 8 April 2019 German Chancellor Angela Merkel plays with a handball given to her by the German Handball Federation's president as she received the German national handball team at the Chancellery in Berlin AFP/Getty 2/50 7 April 2019 People hold candles as they attend a night vigil and prayer at the Amahoro Stadium as part of the 25th Commemoration of the 1994 Genocide, in Kigali, Rwanda. April 7 begins 100 days of mourning for more than 800,000 people who were slaughtered in a genocide that shocked the world, a quarter of a century on from the day it began AFP/Getty 3/50 6 April 2019 A portion of the field of 1,500 participants begins the trek to the highest point over the Florida Keys Overseas Highway's longest span during the Seven Mile Bridge Run Saturday. The event features entrants running a course over the convergence of the Atlantic Ocean and Gulf of Mexico and helps to raise funds for local youth athletic programs AFP/Getty 4/50 5 April 2019 A refugee father and son lie on railway tracks to prevent a train from leaving a station during a protest in Athens, Greece. Dozens of migrants staged a protest in Athens central train station disrupting all railway services in the hope they will be transported to the Greek border and join other refugees attempting to follow a 2016 migration route towards northern Europe Getty 5/50 4 April 2019 Security agents and police officers hold back migrants during the evacuation of a makeshift camp at Porte de la Chapelle, in the north of Paris. More than 300 migrants and refugees were evacuated on early April 4 from a makeshift camp to accomodation structures AFP/Getty 6/50 3 April 2019 An Alexandra township resident gestures and they part is clashes with the Johannesburg Metro Police, South Africa during a total shutdown of the township due to protest against the lack of service delivery or basic necessities such as access to water and electricity, housing difficulties and lack of public road maintenance AFP/Getty 7/50 2 April 2019 Children eat next to the debris of damaged homes at Purainiya village in Nepal's southern Bara district near Birgunj, following a rare spring storm. The freak storm tore down houses and overturned cars and trucks as it swept across southern Nepal killing at least 27 people and leaving more than 600 injured AFP/Getty 8/50 1 April 2019 A forensic expert works next to the remains of a small plane that crashed near Erzhausen, Germany. Natalia Fileva, chairwoman and co-owner of Russia's second largest airline S7, died when a private jet she was in crashed near Frankfurt on Sunday, the company said Reuters 9/50 31 March 2019 Ukrainian comic actor and presidential candidate Volodymyr Zelenskiy delivers a speech following the announcement of the first exit poll in a presidential election at his campaign headquarters in Kiev, Ukraine Reuters 10/50 30 March 2019 Catalan pro-independence protesters throw rocks during a counter-demonstration against a protest called by Spanish far-right party Vox against the Catalan independence push in Barcelona. Polls suggest Vox, which campaigns against illegal immigration and "radical" feminism, will become the first far-right party to win seats in the Spanish parliament since the late 70s and could emerge as a kingmaker in Spain's increasingly fragmented political landscape AFP/Getty 11/50 29 March 2019 Protests against President Abdelaziz Bouteflika continue in Algeria despite the announcement on 11 March that he will not run for a fifth Presidential term and postponement of presidential elections previously scheduled for 18 April 2019 until further notice EPA 12/50 28 March 2019 Firefighters on ladders work to extinguish a blaze in an office building in Dhaka after a huge fire tore through it, killing at least five people with many others feared trapped in the latest major fire to hit the Bangladesh capital AFP/Getty 13/50 27 March 2019...

Israel police mistakenly arrest Palestinian man for writing ‘good morning’ on Facebook
Facebook translates 'good morning' into 'attack them', leading to arrest

theguardian.com · 2017

Palestinian man questioned by Israeli police after embarrassing mistranslation of caption under photo of him leaning against bulldozer

Facebook has apologised after an error in its machine-translation service saw Israeli police arrest a Palestinian man for posting “good morning” on his social media profile.

The man, a construction worker in the West Bank settlement of Beitar Illit, near Jerusalem, posted a picture of himself leaning against a bulldozer with the caption “يصبحهم”, or “yusbihuhum”, which translates as “good morning”.

But Facebook’s artificial intelligence-powered translation service, which it built after parting ways with Microsoft’s Bing translation in 2016, instead translated the word into “hurt them” in English or “attack them” in Hebrew.

Police officers arrested the man later that day, according to Israeli newspaper Haaretz, after they were notified of the post. They questioned him for several hours, suspicious he was planning to use the pictured bulldozer in a vehicle attack, before realising their mistake. At no point before his arrest did any Arabic-speaking officer read the actual post.

Facebook said it is looking into the issue, and in a statement to Gizmodo, added: “Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted.

“Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologise to him and his family for the mistake and the disruption this caused.”

Arabic is considered particularly difficult for many machine translation services due to the large number of different dialects in use around the world, on top of Modern Standard Arabic, the international form of the language.

The Israeli Defence Force has been open about monitoring the social media accounts of Palestinians, looking for “lone-wolf” attackers who might otherwise slip through the net. It reportedly does so automatically, using algorithms to look for terms such as “sword of Allah”.

Machine translation mistakes are a regular occurrence for anyone using AI to translate languages, particularly ones with little relationship. Earlier this month, Chinese social network WeChat apologised after its own machine translation system translated a neutral phrase meaning “black foreigner” as the n-word.

“When I ran the translator, the n-word came up and I was gobsmacked,” said Ann James, who had been texting back and forth with a friend when the faulty translation appeared....

Facebook translates 'good morning' into 'attack them', leading to arrest
Man Arrested After Facebook Translates Turns Greeting Into Threat

digitaltrends.com · 2017

Share

Facebook’s auto-translate feature allows users to connect beyond language barriers but one incorrect translation of a simple “good morning” proved to be a bit of a disaster for one Palestinian man. When the construction worker posted a picture of his work site, the Facebook translate feature reportedly turned the Arabic “good morning” into a Hebrew “attack them” and an English “hurt them,” resulting in the man’s arrest.

According to Israel newspaper Haaretz, the man was arrested when Israeli police spotted the auto-translate message last week, which was accompanied by a photo of the man leaning against a bulldozer enjoying what appears to be a morning coffee and cigarette at a West Bank construction site near Jerusalem. The confusion came from the system misidentifying a similar Arabic word which means “to hurt.”

The incorrect translation flagged the post, notifying local authorities who also use algorithms to flag potential threats. The police responded because of both the translation and the image. According to police, bulldozers have been used in terrorist attacks before as hit-and-run vehicles.

After questioning and after an Arabic-speaking office read the original post, the police realized the error and the man was released after a few hours.

In a statement, Facebook apologized for the error. The company says that, while the auto-translate algorithms improve every day, misinterpretations happen occasionally. Facebook said it is working to prevent the error from happening again.

Artificial intelligence is behind Facebook’s translation feature — when the company switched entirely to its own system last year, the software handled around 2 billion translations a day in 40 languages. Additional options allow users to report bad translations and rate translated text.

The most popular social media network uses a number of different algorithms to monitor posts, but a string of recent events has the platform promising more human oversight to the process in a number of different areas. A set of ad-targeting metrics for hate groups and fake news promoted in the trending section are just a few examples of incidents that prompted changes in the platform’s monitoring systems. Earlier in 2017, the company said it would add 3,000 moderators to review posts flagged by users for violating the Community Standards....

Man Arrested After Facebook Translates Turns Greeting Into Threat
Facebook apologizes after wrong translation sees Palestinian man arrested for posting ‘good morning’

theverge.com · 2017

Facebook has apologized after a Palestinian man was arrested by Israeli police for a post saying “good morning” that its automatic-translation service erroneously translated as “attack them” in Hebrew and “hurt them” in English, reports Israeli newspaper Haaretz. The man is a construction worker near Jerusalem, reports The Guardian. He posted a photo of himself last week leaning against a bulldozer with the caption “يصبحهم”, or “yusbihuhum,” which translates as “good morning.”

Police arrested the man after they were notified of the post and were suspicious he was planning a vehicle attack using the bulldozer. He was released hours later after police realized the mistake. Haaretz reports that before his arrest, no Arabic-speaking officer had read the man’s Facebook post.

Facebook is currently investigating the issue and Necip Fazil Ayan, an engineering manager in Facebook’s language technologies group, said in a statement to Gizmodo that though mistakes might happen, the company’s translations are getting better each day. “Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted. We apologize to him and his family for the mistake and the disruption this caused.”

Facebook’s translations are entirely powered by AI, and around 4.5 billion translations are made each day across the social network....

Facebook apologizes after wrong translation sees Palestinian man arrested for posting ‘good morning’
Palestinian Man Arrested After Facebook Auto-Translates 'Good Morning' As 'Attack Them'

gizmodo.com.au · 2017

A Palestinian construction worker was arrested by Israeli police after Facebook incorrectly translated the text of one of his posts. Haaretz reports that the man uploaded a picture from his job at a construction site with the text "good morning" in Arabic. When officers used Facebook's automatic translation service to read the post, the text was mistranslated as "attack them" in Hebrew and "hurt them" in English.

Image: AP

According to Haaretz, Arabic speakers said the "English transliteration used by Facebook is not an actual word in Arabic but could look like the verb 'to hurt' - even though any Arabic speaker could clearly see the transliteration did not match the translation". No Arabic-speaking officers reportedly saw the post prior to the man's arrest. He was released after several hours of questioning.

Facebook originally used Microsoft's translation AI, but began using a proprietary translation software in 2016. When contacted by Gizmodo, Necip Fazil Ayan, an engineering manager in Facebook's language technologies group, provided the following statement:

Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted. Even though our translations are getting better each day, mistakes like these might happen from time to time and we've taken steps to address this particular issue. We apologise to him and his family for the mistake and the disruption this caused.

In the picture uploaded with the text, a man is propped against a bulldozer at a West Bank construction site. This only added to the perceived suspicion, as bulldozers have been used in hit-and-run terrorist attacks. The Israel Defence Forces monitors Palestinians' social media as a way to detect radicalisation or potential terror attacks. The West Bank is largely Israeli controlled, though pockets have joint Israeli-Palestinian control and conflict over contested land is ongoing.

Facebook offers an entire suite of features that it says can help authorities prevent terrorism, from computer vision to keyword monitoring. Still, critics have long warned these tactics can be misused by state powers as a means of targeted surveillance. In this case, a simple misunderstanding could have cost a man his livelihood.

[Haaretz]...

Palestinian Man Arrested After Facebook Auto-Translates 'Good Morning' As 'Attack Them'
Man Arrested After Facebook Translation Error

swordandscale.com · 2017

They say a picture is worth a thousand words, but it only took two of them for a Palestinian man to find himself being placed under arrest after a Facebook translation blunder.

A construction worker for the West Bank settlement of Beitar Illit took a photo of himself posing with a bulldozer. Along with the photo, the construction worker captioned his selfie with “good morning.” Facebook’s translator service, however, misinterpreted the sentiment and instead read the phrase as stating “hurt them” in English and “attack them” in Hebrew.

Within hours of the post, Israeli police placed the man under arrest after receiving reports. None of the Arabic-speaking officers read the post prior to the man’s arrest. Police interrogated the man, believing he was plotting to use the bulldozer in the photo as part of a strategic lone-wolf attack against Israelis. Police soon realized the mistake and released the construction worker, according to Haaretz, an Israeli newspaper.

The tech blog Gizmodo reached out to Facebook about the blunder. Necip Fazil Ayan, who is purported to be an engineering manager for Facebook’s language technologies group, responded with the following statement:

“Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted. Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologize to him and his family for the mistake and the disruption this caused.”

The Guardian explains that Arabic can be particularly difficult to translate due to the sheer number of dialects that are used throughout the world. AI translation mistakes are common, especially when asked to translate specific words in unrelated languages....

Man Arrested After Facebook Translation Error
Facebook's auto translation AI fail leads to a nightmare for a Palestinian man

mashable.com · 2017

Bad translations often make good comedy, but in the case of this Palestinian man, they marked the start of a nightmare.

Israeli police arrested the man, a construction worker in a West Bank settlement near Jerusalem, after he posted on Facebook a picture of himself leaning against a bulldozer with the caption reading "يصبحهم" or "Good Morning" in Arabic.

Facebook's automatic translation service, which uses its own AI system, translated the post into "attack them" in Hebrew, or "hurt them" in English:

In predictive policing news: Israel arrested a Palestinian for posting a pic on Facebook next to a bulldozer saying “good morning” just because Facebook incorrectly translated the Arabic phrase to mean “hurt them”. Full article on Haaretz - https://t.co/kk1A51BcUy pic.twitter.com/BUOVONWIrc — Asaf Lubin (@AsafLubin) October 23, 2017

For this reason, the Israeli Judea and Samaria District police detained the man ad questioned him for several hours on suspicion he's going to use the bulldozer for a hit-and-run attack, according to Haaretz.

During this whole process, no Arabic speaker was consulted to double-check on the Facebook's automatic translation.

After realising their mistake, police released the man, Haaretz reported.

Facebook told Mashable it was looking into the issue. In a statement to Gizmodo, Facebook's engineering manager Necip Fazil Ayan said:...

Facebook's auto translation AI fail leads to a nightmare for a Palestinian man
Israel: Botched Facebook translation lands man in lock-up; ‘Good Morning’ turns ‘Attack Them’

newsx.com · 2017

Palestinian builder Halawim Halawi who posted 'Good Morning' on his Facebook account which awarded him a few hours in the lockup. Facebook's automatic translation software which interpreted the post to mean "attack them" in Hebrew and "hurt them" in English.

Israeli police arrested a man from West Bank after he wrote ‘good morning’ on his Facebook in Arabic. The Facebook translate software mistranslated it as “attack them” and the man was busted by Israeli police. This is something which occurs now and then in this age of internet and social media boom as translations by software often lead to miscommunications. Such was the case with Palestinian builder Halawim Halawi who posted ‘Good Morning’ on his Facebook account but the greeting did not go well and he was arrested by police shortly after.

The culprit, in this case, was Facebook’s automatic translation software which interpreted the post to mean “attack them” in Hebrew and “hurt them” in English. Understandably, the Israeli police took note of the Facebook post in which the man is seen standing next to a bulldozer and the ensuing interpretation meant some lock-up time for the early riser who, like many others, likes to send good morning messages on social media. As per reports, such heavy vehicles had been earlier used in terrorist attacks so they were suspicious and acted immediately and arrested the man.

Later, the police realised that they had made a mistake and released the man after a few hours of questioning. The questioning would surely have been interesting as the cops would have tried to make heads and tails of the ‘threat’ for 4 hours and yet not come to a conclusion that the person in question was a national security threat. Police spokeswoman Luba Samri told a news agency that “a few days ago, a Palestinian was detained for questioning on suspicion of incitement through his Facebook page.” She said he was “immediately released” after the suspicions turned out to be false.

For all the latest World News, download NewsX App

Read More...

Israel: Botched Facebook translation lands man in lock-up; ‘Good Morning’ turns ‘Attack Them’
A ‘Good Morning’ Reportedly Gets Man Arrested After Facebook Translate Error

supernewsworld.com · 2017

Digital Trends has reported that Facebook’s auto-translate feature allows users to connect beyond language barriers but one incorrect translation of a simple “good morning” proved to be a bit of a disaster for one Palestinian man. When the construction worker posted a picture of his work site, the Facebook translate feature reportedly turned the Arabic “good morning” into a Hebrew “attack them” and an English “hurt them,” resulting in the man’s arrest.According to Israel newspaper Haaretz, the man was arrested when Israeli police spotted the auto-translate message last week, which was accompanied by a photo of the man leaning against a bulldozer enjoying what appears to be a morning coffee and cigarette at a West Bank construction site near Jerusalem. The confusion came from the system misidentifying a similar Arabic word which means “to hurt.” https://www.digitaltrends.com/social-... http://www.wochit.com This video was produced by YT Wochit News using http://wochit.com...

A ‘Good Morning’ Reportedly Gets Man Arrested After Facebook Translate Error
Man arrested for saying 'good morning' because online translator made a mistake

metro.co.uk · 2017

A construction worker was arrested after he posted ‘Good morning’ on Facebook, because the automatic translation got it wrong.

The Palestinian man had posted a photo at work with a cigarette and a coffee, adding a greeting to his followers in Arabic.

But the online translator misread what he said, translating it as ‘Hurt them’.

Israeli police arrested him on suspicion of inciting violence, believing he was urging people to use a bulldozer to attack people, the Times of Israel reports.

Nobody who spoke Arabic had read the post before he was interrogated.

Israeli security forces arrested him (File picture: Getty)

The automatic translator got mixed up because there is only one letter different in the phrases ‘good morning to you all’ and ‘hurt them’ in Arabic.

Bulldozers have been used in terror attacks in the past, so police thought he was promoting violence.

In fact it was a totally innocent and good natured photo from Beitar Illit settlement on October 15.

Advertisement

Advertisement

The man has since deleted his Facebook post and did not want to be identified, the Times reported.

MORE: Nursery worker died after accidentally overdosing on medication for mystery pain

MORE: Couple find huge stash of marijuana hidden inside their Amazon order...

Man arrested for saying 'good morning' because online translator made a mistake
Israel arrested a Palestinian after Facebook translated “Good morning” as “Attack them”

qz.com · 2017

Facebook says it automatically translates about 4.5 billion posts each day. With new research and advancements, the technology keeps improving, but it’s not perfect—and its flaws can have serious consequences.

Last week, a Palestinian man working at a construction site in an Israeli settlement in the West Bank, posted on Facebook a photo of himself in front of a bulldozer, with the words “Good Morning” in Arabic. The platform’s automatic translation mistook the words for “attack them” in Hebrew, and “hurt them” in English.

Here’s an explanation from the Israeli daily newspaper Haaretz, which first reported the incident, about the nature of the mistake:

Arabic speakers explained that English transliteration used by Facebook is not an actual word in Arabic but could look like the verb “to hurt” – even though any Arabic speaker could clearly see the transliteration did not match the translation.

Google Translate also mistranslates the phrase, turning it into “become them,” likely because the literal translation of “morning” in Arabic is “day becoming.”

The Jerusalem Post reports that Israeli police got multiple complaints from civilians about the post, and that they acted upon it because of previous terrorist attacks using a bulldozer. Reportedly, no officer who spoke Arabic saw the post before the arrest was made. The police released the man from custody after several hours, realizing the mistake.

Facebook introduced its translation tool in 2011. Then, it used Microsoft’s Bing to translate. In late 2015 the platform completed a shift to its own AI translation technology. In 2016, users started being able to have their statuses automatically translated, visible to everyone in the language they post in. In August 2017, the platform fully transitioned from phrase-based translation to the more accurate neural network system. The company is developing a more sophisticated neural technology, which it says is nine times faster than its competitors.

Recently, another case of mistranslation by Facebook resulted in “Mexicanitos,” meant to be a term of endearment, being posted as the slur “wetbacks.” The mother from El Paso who posted the photo of her kids, calling them “Mexicanitos,” complained publicly....

Israel arrested a Palestinian after Facebook translated “Good morning” as “Attack them”
Facebook Apologizes After Translation Fail Leads to False Arrest

dailydot.com · 2017

No translation service is perfect, but you wouldn’t expect it to get you arrested.

Facebook admitted that an error in its artificial intelligence-based auto-translation feature led to the arrest of a Palestinian man who posted “good morning” on his social profile.

The man, a construction worker in the West Bank near Jerusalem, posted a picture of himself leaning against a bulldozer with the caption “??????”, or “yusbihuhum,” which means “good morning” in Arabic.

In predictive policing news: Israel arrested a Palestinian for posting a pic on Facebook next to a bulldozer saying “good morning” just because Facebook incorrectly translated the Arabic phrase to mean “hurt them”. Full article on Haaretz – https://t.co/kk1A51BcUy pic.twitter.com/BUOVONWIrc — Asaf Lubin (@AsafLubin) October 23, 2017

Facebook’s translator replaced his pleasant greeting with “hurt them” in English or what translates to “attack them” in Hebrew.

No Arabic-speaking authorities got a look at his post in time and police arrested the man later that day on suspicion that he was planning an attack using the bulldozer, a vehicle that had previously been used in hit-and-run terrorist attacks. He was released after several hours of questioning.

Arabic speakers tell Haaretz that the transliteration produced by Facebook isn’t a real Arabic word, but looks a lot like “to hurt.” A native Arabic speaker would reportedly have spotted the error.

Facebook apologized and said it would investigate the problem in a statement to Gizmodo.

“Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted. Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologize to him and his family for the mistake and the disruption this caused.”

H/T the Guardian...

Facebook Apologizes After Translation Fail Leads to False Arrest
Facebook translates 'good morning' into 'attack them', leading to arrest

theguardian.com · 2017

Palestinian man questioned by Israeli police after embarrassing mistranslation of caption under photo of him leaning against bulldozer

Facebook has apologised after an error in its machine-translation service saw Israeli police arrest a Palestinian man for posting “good morning” on his social media profile.

The man, a construction worker in the West Bank settlement of Beitar Illit, near Jerusalem, posted a picture of himself leaning against a bulldozer with the caption “يصبحهم”, or “yusbihuhum”, which translates as “good morning”.

But Facebook’s artificial intelligence-powered translation service, which it built after parting ways with Microsoft’s Bing translation in 2016, instead translated the word into “hurt them” in English or “attack them” in Hebrew.

Police officers arrested the man later that day, according to Israeli newspaper Haaretz, after they were notified of the post. They questioned him for several hours, suspicious he was planning to use the pictured bulldozer in a vehicle attack, before realising their mistake. At no point before his arrest did any Arabic-speaking officer read the actual post.

Facebook said it is looking into the issue, and in a statement to Gizmodo, added: “Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted.

“Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologise to him and his family for the mistake and the disruption this caused.”

Arabic is considered particularly difficult for many machine translation services due to the large number of different dialects in use around the world, on top of Modern Standard Arabic, the international form of the language.

The Israeli Defence Force has been open about monitoring the social media accounts of Palestinians, looking for “lone-wolf” attackers who might otherwise slip through the net. It reportedly does so automatically, using algorithms to look for terms such as “sword of Allah”.

Machine translation mistakes are a regular occurrence for anyone using AI to translate languages, particularly ones with little relationship. Earlier this month, Chinese social network WeChat apologised after its own machine translation system translated a neutral phrase meaning “black foreigner” as the n-word.

“When I ran the translator, the n-word came up and I was gobsmacked,” said Ann James, who had been texting back and forth with a friend when the faulty translation appeared....

Facebook translates 'good morning' into 'attack them', leading to arrest
Facebook apologizes after wrong translation sees Palestinian man arrested

news.ycombinator.com · 2017

Hacker news discussion of Facebook's apology....

Facebook apologizes after wrong translation sees Palestinian man arrested
Faulty Facebook translation leads to Palestinian man’s arrest

tomedes.com · 2017

It was only last week that we discussed the flaws of machine translation. In that case, the focus of the article was on the WeChat messaging app’s racial slurs. Now, a new story has once more highlighted the dangers posed by the machines that we rely on to translate for us.

Arrest in Israel

The case in question concerns a Palestinian gentleman working in Israel. The man works in the construction industry near Jerusalem. He recently posted a photograph of himself at work, leaning against a bulldozer. He added the caption “يصبحهم”, or “yusbihuhum,” which means “good morning.”

Unfortunately, this seemingly innocent act triggered a great deal of trouble for this individual. Facebook’s artificial intelligence (AI) powered automatic translation system, which handles some 4.5 billion translations per day, made a mistake. Instead of translating the man’s innocent greeting correctly, the system translated it as “attack them” in Hebrew and “hurt them” in English. Local police put the translated comment together with the bulldozer in the picture and promptly arrested the poor construction worker.

Human translation lacking

Sadly, no Arabic translation was undertaken before the arrest. Nor were any Arabic-speaking officers consulted prior to the gentleman being hauled in for questioning. The suspicion that the gentleman was planning some form of bulldozer attack caused police to react first and ask questions later. Of course, once they did begin asking questions, the truth quickly became apparent.

The incident serves to highlight once more the importance of human translation rather than reliance on machines. In this case, the issue was resolved within a few hours when the erroneous translation was discovered. However, one can’t help but wonder how many other instances of fault Facebook translation are taking place each and every day, as users rely on artificial intelligence to communicate in other languages.

Facebook’s apology

Facebook has, quite rightly, apologised for making such a mistake. Necip Fazil Ayan, an engineering manager in Facebook’s language technologies group, commented,

“Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted.

“Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologise to him and his family for the mistake and the disruption this caused.”

Machine translation

The translation error occurred in part due to the difficulties that machines face when it comes to translating Arabic. Facebook’s system is far from alone in finding Arabic a hard language to translate. As well as the internationally used Modern Standard Arabic, the language has a large number of different dialects. This provides machines with a level of complexity that they don’t often face when working with other languages.

Even so, the error shows how far machine translation still has to go before we can rely on it. If a simple phrase such as “good morning” can serve to flummox Facebook’s translation system so utterly, despite the enormous funds available to Facebook to invest in that system, one shudders to think of how poorly more complex chunks of text are being translated.

Over the past few years, we’ve heard about numerous breakthroughs in the field of machine translation, from Google’s neural networks translating without transcribing courtesy of deep learning techniques, to Facebook’s much publicised shift to a system of entirely neural machine translation. So why is it that machine translation is still so flawed compared to human translation?

Google recently announced that its Deep Mind project had made a leap forward in AI research. The AlphaGo Zero project saw Deep Mind achieve superhuman abilities at the game Go after being given a board and a set of instructions with which to teach itself. Whereas the previous project – AlphaGo – required thousands of hours of learning from human players before it was able to beat a human champion, AlphaGo Zero taught itself to do so in just a few days, defeating the original AlphaGo by 100 games to 0. The machine even taught itself Go techniques and strategies that humans have never thought of, despite the game being over a thousand years old.

Perhaps this new approach to machine learning will finally be the masterstroke that leads to machines being able to translate as proficiently as humans – but, then again, we’ve heard such proclamations many times before and humans still have the edge over machines when it comes to professional translation!

Final thoughts

Will the latest advances in AI lead to machine translation being perfected? Or will we continue to see machine translation mistakes leading to confusion and, in extreme cases, arrests? Share your thoughts by leaving a comment below....

Faulty Facebook translation leads to Palestinian man’s arrest
This Man Was Arrested For Posting 'Good Morning' On Facebook

thealternativedaily.com · 2017

Israeli police accidentally arrested a Palestinian man after he posted “Good morning” in Arabic on Facebook, together with a photo of himself at work.

The man was working at a building site on a West Bank settlement near Jerusalem and posed with a photo of a bulldozer in the post. Facebook’s automatic language translation software translated the post as “attack them” in Hebrew and “hurt them” in English.

According to Israeli paper Haaretz, no Arabic-speaking police officer read the post before the man was arrested. After questioning, the police realized their mistake, and released the man a few hours later. At that point, the police agreed that the correct translation was “good morning.”

Facebook also quickly apologized for the mistake, Gizmodo reported. “Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted,” said Necip Fazil Ayan, an engineering manager in Facebook’s language technologies group. “We apologize to him and his family for the mistake and the disruption this caused.”

Facebook began using its AI translation system in 2015, with translations available since 2011. In this case, Arabic speakers explained to Haaretz that the English transliteration that Facebook used wasn’t a real word in Arabic, but that it could look like the verb “to hurt,” though “any Arabic speaker could clearly see the transliteration did not match the translation.”

Facebook starts using its own translation system

Since last year, Facebook has been translating two billion text snippets per day. The application can translate 40 languages in 1,800 directions (such as French to English, or Japanese to Spanish). According to TechCrunch, some 800 million users — nearly half of the total number of people using Facebook — see at least some translations every month.

The advantage for Facebook in using its own translation system is that it can have more control over people’s news feeds, by understanding the meaning behind text and images. For example, if Facebook understands that a post is looking for recommendations of sites to visit in Rome, no matter if a post is in Indonesian, Arabic or Mandarin, it can suggest friends to ask for that recommendation.

Now, when Facebook is confident that its translation is correct, it automatically shows that translation, with an option to “view original,”

However, AI translation depends on learning from the most common uses of words, and most bi- or multi-lingual speakers find that Google and Facebook translations are very lacking.

“To have a machine being as smart as a human and holding up a conversation — where you actually have a meaningful counterpart — that’s still pretty science fiction at the moment,” said Philip Koehn, a Johns Hopkins University lecturer who has studied machine translation.

So, were police right to intervene and arrest that Facebook user? Let us know in the comments below!

— Tamara Pearson...

This Man Was Arrested For Posting 'Good Morning' On Facebook
Palestinian Arrested After Facebook Translation Error

forward.com · 2017

A Palestinian man was mistakenly arrested in Israel after Facebook’s translation software turned his “good morning” photo caption into a threat.

The man, who works in a West Bank settlement, posted a photo of himself next to a bulldozer with the message “good morning” in Arabic. But Facebook’s proprietary translation algorithm translated it into Hebrew as “attack them.”

Local police officers were notified of the post and arrested the man. They were especially concerned that he was posing next to a bulldozer, which has been used by Palestinians in hit-and-run terror attacks.

Israeli security frequently monitors Palestinians’ social media accounts to detect possible terror threats.

No Arabic-speaking officer read the post before the man’s arrest, Haaretz reported. The man was soon freed after the police admitted a mistake was made.

“Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted,” a Facebook spokesperson told Gizmodo. “Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologize to him and his family for the mistake and the disruption this caused.”

Contact Aiden Pink at pink@forward.com or on Twitter, @aidenpink...

Palestinian Arrested After Facebook Translation Error
How a Facebook automatic translation error got an innocent man arrested

translateplus.com · 2017

Some automatic translation caused mistakes are harmless enough; others not so much. The occasional language faux-pas can be good for a laugh but you never know when a seemingly harmless mishap can cause real damage. The latest example of translation gone wrong comes from Facebook after its AI-powered automatic translation system got a Palestinian construction worker arrested for suspected terrorist activities.

Whoops.

Which acts as another reminder of how important it is to have machine translation checked and verified by a professional.

Facebook automatic translation gets man arrested

The unfortunate Palestinian man in this story was guilty of little more than posting a picture of himself on Facebook, leaning against a bulldozer. His caption simply read “??????” or “Good Morning” in Arabic and there was no hint of anything that should alert the authorities about a potential terrorist attack.

In predictive policing news: Israel arrested a Palestinian for posting a pic on Facebook next to a bulldozer saying “good morning” just because Facebook incorrectly translated the Arabic phrase to mean “hurt them”. Full article on Haaretz – https://t.co/kk1A51BcUy pic.twitter.com/BUOVONWIrc — Asaf Lubin (@AsafLubin) October 23, 2017

That was until Facebook’s automatic translation AI interfered with things. Instead of translating the caption into “good morning”, Facebook managed to come up with “attack them” in Hebrew and “hurt them” in English. Which raised the alarm with authorities who suspected the man in question was planning to use the bulldozer in a hit-and-run terrorist attack.

Bulldozers have been used in the past for such attacks.

Man released, lesson learned

Once the automatic translation error was pinpointed, the man was released by police. He has since deleted the original post and neglected to talk to the press about his Facebook translation ordeal. Facebook, on the other hand, was somewhat obliged to make a statement after inadvertently getting an innocent man arrested:

“Unfortunately, our translation systems made an error last week that misinterpreted what this individual posted. Even though our translations are getting better each day, mistakes like these might happen from time to time and we’ve taken steps to address this particular issue. We apologize to him and his family for the mistake and the disruption this caused.” – Necip Fazil Ayan, Facebook’s engineering manager.

Sadly, there’s little the man arrested over the mixup could have done differently. Translating the innocent phrase “good morning” into two separate versions that triggered terrorism alarm bells is a pretty drastic mistake. Which highlights the importance of having machine translation verified by humans before they go live.

Above all, this reminds us that a simple automatic translation mixup can have serious consequences. It’s not the mistake itself that’s a problem but the repercussions that matter. In this case a man was arrested under anti-terror laws and promptly released. However, not all translation mistakes are reversible, even if we assume this instance hasn’t had lasting effects. In more drastic cases businesses have lost millions and people have even lost their lives due to simple but costly automatic translation mistakes.

As always, mistakes will happen. This is why machine translation should always be checked and verified by professionals – just like human translations are checked by other pros before going live. Reducing the frequency and consequences of translation errors is vital, particularly in fields like law and medicine where people’s lives are at stake. Regarding Facebook’s latest automatic translation error, the network is putting itself in the legal spotlight itself if more people get inadvertently arrested – or worse – due to its AI glitches.

(Visited 177 times, 1 visits today)...

How a Facebook automatic translation error got an innocent man arrested
Man arrested after ‘good morning’ post was mistranslated by Facebook as ‘attack them’

forum.facepunch.com · 2018

[QUOTE]Israeli police arrested a Palestinian man after his “good morning” post was translated by Facebook as “attack them.” Social media monitoring combined with Facebook’s AI mistranslating a post led to a man being wrongfully arrested for “suspicion of incitement.” A smiling Palestinian construction worker posted a photo of himself leaning against a bulldozer and holding a cup of coffee and a cigarette. He posted the photo on Facebook along with “good morning” in Arabic. Israeli police, relying on Facebook’s translation service, believed the post said “attack them.” Haaretz reported, “The automatic translation service offered by Facebook uses its own proprietary algorithms. It translated ‘good morning’ as ‘attack them’ in Hebrew and ‘hurt them’ in English.” ... Israeli police did not consult any Arabic-speaking officer, but instead relied completely on Facebook’s translation. As a result, the Palestinian man was arrested. ... The Palestinian construction worker deleted his “good morning” post after he was arrested and then released[/QUOTE] [URL="https://www.csoonline.com/article/3234644/security/man-arrested-after-good-morning-post-was-mistranslated-by-facebook-as-attack-them.html"]Source[/URL]

What an idiotic knee-jerk arrest. Why would you arrest someone before getting an accurate translation?

[QUOTE=Natrox;52811186]What an idiotic knee-jerk arrest. Why would you arrest someone before getting an accurate translation?[/QUOTE] Because they were looking for an excuse to arrest a palestinian.

Didn't know Blizzerd was working at Facebook now :v:

I don't think anyone should rely on translation engines to replace human translations. Google translate can be pretty shit sometimes as well, going totally off.

[QUOTE=Electrocuter;52811197]Didn't know Blizzerd was working at Facebook now :v:[/QUOTE] Australia will never forget

[QUOTE=Scratch.;52811249]Australia will never forget[/QUOTE] Context?

[QUOTE=Ignhelper;52811220]I don't think anyone should rely on translation engines to replace human translations. Google translate can be pretty shit sometimes as well, going totally off.[/QUOTE] It's not so much a problem of not relying on AI translation systems, but rather knowing their uncertainty about a certain translation. For example, Google's translate in particular uses something called [URL=https://research.googleblog.com/2016/11/zero-shot-translation-with-googles.html]"Zero-Shot" translation[/URL] which essentially allows "translation" of never before seen text. This means that there's a high chance that you can get a bad translation for something the system doesn't know about, and then you get hilarious side-effects like [URL=https://www.youtube.com/watch?v=3-rfBsWmo0M]these[/URL], or horrible like this arrest.

[QUOTE=Ignhelper;52811220]I don't think anyone should rely on translation engines to replace human translations. Google translate can be pretty shit sometimes as well, going totally off.[/QUOTE] Eventually these translations may get so good that machine translations *are* the real meaning of a term, legally. Imagine what happens when you post something in a non-native language and the translation translates much better, but still strips sarcasm and context or makes a rare error. Getting stuck/unhirable/arrested because of some weird translation error like that would be the ultimate bureaucratic nightmare, namely because there would be no human bureaucracy even making the decision to begin with. (and this is our likely future)

[QUOTE=SlickBlade;52811340]Context?[/QUOTE] Everyone in Australia is an elephant.

I already knew about the one like that in [URL="https://kasko.helifreak.club/image/a7KMR1m_700b_v1.jpg"]Brazilian and Indonesian[/URL] but I didn't know there was something similar in Arabic.

[QUOTE=Crimor;52811190]Because they were looking for an excuse to arrest a palestinian.[/QUOTE] disturbing shit, even if he DID post 'attack them' in the context of a casual photo how could that possibly be probable cause to arrest someone? makes this statement sketchy as fuck about how many of these 'prevented attacks' are real: [QUOTE]Like most government agencies, Israel’s security services conduct social media monitoring; it led to the arrests of 400 Palestinians suspected of planning attacks in April. Israel claimed its social media monitoring had stopped a total 2,200 Palestinians who were planning attacks.[/QUOTE]

[QUOTE=Ignhelper;52811220]I don't think anyone should rely on translation engines to replace human translations. Google translate can be pretty shit sometimes as well, going totally off.[/QUOTE] [video=youtube;PHXkhUJWTXM]http://www.youtube.com/watch?v=PHXkhUJWTXM[/video]...

Man arrested after ‘good morning’ post was mistranslated by Facebook as ‘attack them’
Mixed Messages? The Limits of Automated Social Media Content Analysis

cdt.org · 2018

Governments and companies are turning to automated tools to make sense of what people post on social

media. Policymakers routinely call for social media companies to identify and take down hate speech,

terrorist propaganda, harassment, fake news or disinformation. Other policy proposals have focused on

mining social media to inform law enforcement and immigration decisions. But these proposals wrongly

assume that automated technology can accomplish on a large scale the kind of nuanced analysis that

humans can do on a small scale. Today’s tools for analyzing social media text have limited ability to parse

the meaning of human communication or detect the intent of the speaker.

A knowledge gap exists between data scientists studying natural language processing (NLP) and

policymakers advocating for wide adoption of automated social media analysis and moderation.

Policymakers must understand the capabilities and limits of NLP before endorsing or adopting automated

content analysis tools, particularly for making decisions that affect fundamental rights or access to

government benefits. Without proper safeguards, these tools can facilitate overbroad censorship and

biased enforcement of laws or terms of service.

This paper draws on existing NLP research to explain the capabilities and limitations of text classifiers for

social media posts and other online content. It is aimed at helping researchers and technical experts

address the gaps in policymakers’ knowledge about what is possible with automated text analysis. We

provide an overview of how NLP classifiers work and identify five key limitations of these tools that must

be communicated to policymakers: (1) Natural language processing tools perform best when they are

trained and applied in specific domains, and cannot necessarily be applied with the same reliability across

different contexts; (2) Decisions based on automated social media content analysis risk further

marginalizing and disproportionately censoring groups that already face discrimination. NLP tools can

amplify social bias reflected in language and are likely to have lower accuracy for minority groups who

are underrepresented in training data; (3) Accurate text classification requires clear, consistent definitions

of the type of speech to be identified. Policy debates around content moderation and social media mining

tend to lack such precise definitions; (4) The accuracy and intercoder reliability challenges documented in

NLP studies warn against widespread application of the tools for consequential decision-making; and (5)

Text filters remain easy to evade and fall far short of humans’ ability to parse meaning from text. Human review of flagged content remains essential for avoiding over-censorship. The paper concludes with

recommendations for NLP researchers to bridge the knowledge gap between technical experts and

policymakers....

Mixed Messages? The Limits of Automated Social Media Content Analysis
A Google translation isn't enough evidence to send someone to jail, finds the judge in a narcotics case (GOOG, GOOGL)

finance.yahoo.com · 2018

Greg Sandoval/Business Insider

A man from Mexico was pulled over on the highway in Kansas and was suspected of carrying drugs.

With the help of Google Translate, the Kansas Highway trooper asked the man in Spanish if he can search his car. The suspect says he didn't understand the request when he responded yes.

The question the court was trying to answer is whether Google Translate is a reliable enough interpreter to justify sending a man to prison — and the answer was no.

As anyone who has traveled overseas can tell you, Google Translate is an important tool that can bridge many language barriers.

But while the software’s reliability at offering accurate translations has greatly improved over time, the technology remains far from perfect. US District Judge Carlos Murguia found on June 4 that the software's ability to interpret accurately isn't dependable enough to be the deciding factor in sending someone to prison. Quartz first reported on this story.

According to court documents, Ryan Wolting, a Kansas Highway Patrol trooper stopped Omar Cruz-Zamora last October for driving with a suspended registration. Prosecutors say Wolting began to use Google Translate to communicate after realizing that Cruz-Zamora spoke little English. The trooper asked Cruz-Zamora if he could search his car and Cruz-Zamora responded yes. The documents say the officer then found 14 pounds of meth and cocaine.

Later, Cruz-Zamora’s attorneys filed a motion to suppress the evidence, arguing that the defendant didn’t understand Wolting’s request to search his car or that he had the right to decline the search.

Wolting testified that he keyed into Google Translate the question “Can I search your car?” or “Can I search the car?” The problem was that the translation offered was “¿Puedo buscar el auto?” In Spanish this can be interpreted as “Can I find the car?”

According to a recording made by the police car's camera, there were multiple times when Wolting’s questions produced “nonsensical translations.” Two professional interpreters testified that Google Translate can be used for literal translations but should never be used to “translate full conversations.”

The judge wrote in his decision that "the court does not believe it is reasonable to rely on the service to obtain consent" and granted the defendant’s request to suppress the evidence.

Translation software is an essential tool for travelers, hotel workers, waiters and waitresses who work in cities that see a lot of tourists. But they should realize that the technology is sometimes dramatically incorrect. In October, Israeli police arrested a Palestinian man for posting a photo of himself to his Facebook page standing next to a tractor and writing the words “good morning.”

The social network’s translation tools goofed and translated the words in English as “hurt them”and in Hebrew as “attack them.” Authorities believed the man may have planned to use the tractor in a attack. The man was eventually released and Facebook apologized but there's no mistaking that AI-generated translations still have a long ways to go.

NOW WATCH: We tried gaming on the Samsung CHG90 ultrawide gaming monitor

See Also:

SEE ALSO: We put Google's new language-translation headphones to the test with 10 different languages — here’s how they did...

A Google translation isn't enough evidence to send someone to jail, finds the judge in a narcotics case (GOOG, GOOGL)