Associated Incidents
Beth Hyland thought she had met the love of her life on Tinder.
In reality, the Michigan-based administrative assistant had been manipulated by an online scam artist who posed as a French man named 'Richard', used deepfake video on Skype calls, and posted photos of another man to pull off his con.
A 'deepfake' is manipulated video or audio made using artificial intelligence (AI) to look and sound real. They are often difficult to detect without specialised tools.
In a matter of months, Hyland, 53, had taken out loans totalling $26,000 (Dh95,497), sent 'Richard' the money, and fallen prey to a classic case of romance baiting.
A projected eight million deepfakes will be shared worldwide in 2025, up from 500,000 in 2023, says the British government.
About a fifth of those will be part of romance scams, according to a January report from cyber firm McAfee.
"It's like grieving a death," Hyland said. "When I saw him on video, it was the same as the pictures he had been sending me. He looked a little fuzzy, but I didn't know about deepfakes."
Manipulation and lies
Hyland lives in Portage, about 230km west of Detroit, and had been divorced for four years when she began dating again.
She matched on Tinder with a man whose profile seemed to complement hers well.
Now, she says this 'perfect match' was likely orchestrated.
'Richard' said he was born in Paris, but lived in Indiana and worked as a freelance project manager for a construction company that required a lot of travel, including to Qatar.
Months of emotional manipulation, lies, fake photos and AI-doctored Skype calls followed. The scammer pledged his undying love, but had myriad reasons to miss every potential meet-up.
Weeks after they matched, 'Richard' convinced Hyland that he needed her help to pay for a lawyer and a translator in Qatar.
"I told him I was going to take out loans and he started crying, telling me no one's ever loved him like this before," said Hyland in an online interview.
But 'Richard' kept asking for more money and when Hyland eventually told her financial advisor what was happening, he said she was most likely the victim of a romance scam.
"I couldn't believe it, but I couldn't ignore it," said Hyland.
She confronted 'Richard'; he initially denied it all, but then went silent when Hyland asked him to "prove her wrong" and return her money.
Police told Hyland they could not take her case further because there was no "coercion, threat, or force involved", according to a letter from Portage's director of public safety, seen by the Thomson Reuters Foundation.
The office of public safety --- which oversees both the police and fire services --- did not respond to a request for comment.
In an email sent to Hyland after she flagged the scammer's account to Tinder, which was seen by the Thomson Reuters Foundation, the company said it removes users who violate their terms of service or guidelines.
While Tinder said it could not share the outcome of the investigation due to privacy reasons, it said Hyland's report was "evaluated" and "actioned in accordance with our policies".
A Tinder spokesperson said the company has "zero tolerance" of fraudsters and uses AI to root out potential scammers and warn its users. It also offers factsheets on romance scams.
In March, Hyland attended a US Senate committee hearing when a bill was introduced to require dating apps to remove scammers and notify users who interact with fake accounts.
The senator proposing the bill said Hyland's story showed why the legislation was needed.
In general, dating apps do not notify users who have communicated with a scammer once the fraudster's account has been removed or issue alerts about how to avoid being scammed, as required in the proposed new bill.
The US reported more than $4 billion in losses to romance baiting scams in 2023, according to the FBI.
Microsoft, which owns Skype, directed the Thomson Reuters Foundation to blog posts informing users how to prevent romance scams and steps it had taken to tackle AI-generated content, such as adding watermarks to images.
The company did not provide further comment.
Jason Lane-Sellers, a director of fraud and identity at LexisNexis Risk Solutions, said only 7 per cent of scams are reported, with victims often held back by shame.
'AI arms race'
Jorij Abraham, managing director of the Global Anti-Scam Alliance, a Netherlands-based organisation to protect consumers, said humans won't be able to detect manipulated media for long.
"In two or three years, it will be AI against AI," he said.
"[Software exists] that can follow your conversation --- looking at the eyes, if they're blinking --- these are giveaways that something is going on that humans can't see, but software can."
Lane-Sellers from LexisNexis Risk Solutions described it as an AI "arms race" between scammers and anti-fraud companies trying to protect consumers and businesses.
Richard Whittle, an AI expert at Salford Business School in northern England, said he expects future deepfake detection technology will be built in by hardware makers such as Apple, Google, and Microsoft, who can access users' webcams.
Neither Apple nor Google responded to requests for comment on how they protect consumers against deepfakes, or on future product developments.
Abraham said the real challenge was to catch the scammers, who often work in different countries to those they target.
Despite her dead end, Hyland still believes it is good to report scams and help authorities crack down on scammers.
And she wants scam victims to know it is not their fault.
"I've learned terminology... we don't lose (money) or give it away --- it's stolen. We don't fall for scams --- we're manipulated and victimised. "