Skip to Content
logologo
AI Incident Database
Donate
Discover
Submit
  • Welcome to the AIID
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Spatial View
  • Blog
  • AI News Digest
  • Random Incident
  • Sign Up
Discover
Submit
  • Welcome to the AIID
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Spatial View
  • Blog
  • AI News Digest
  • Random Incident
  • Sign Up

Incident 450: Kenyan Data Annotators Reported Psychological Trauma from Labeling Graphic Text for ChatGPT Safety Systems

Description: Sama data annotators in Nairobi, Kenya, reported lasting psychological trauma after labeling graphic sexual and violent text for OpenAI's content-moderation classifiers used in ChatGPT. Workers alleged low pay and inadequate psychological support; Sama disputed aspects of those accounts. The contract ended early, and former workers later petitioned Kenya's Parliament for stronger protections.
Editor Notes: Work began in November 2021; 11/01/2021 approximates the month because the first day is unspecified. Workers were notified of the project's termination in February 2022, with final labeled data delivered in March 2022. TIME published its investigation on 01/18/2023. The Wall Street Journal reported the workers' parliamentary petition on 07/11/2023.

Tools

New ReportNew ResponseDiscoverView History
The OECD AI Incidents and Hazards Monitor (AIM) automatically collects and classifies AI-related incidents and hazards in real time from reputable news sources worldwide.
 

Entities

View all entities
Alleged: OpenAI , Large language model developers and Chatbot developers developed an AI system deployed by OpenAI and Sama, which harmed Sama , Employees of Sama , Data annotators and Sama data annotators in Kenya.
Alleged implicated AI systems: OpenAI large language models , Large language models , ChatGPT , Content moderation systems , AI training data pipelines and OpenAI toxic-content classifiers

Incident Stats

Incident ID
450
Report Count
8
Incident Date
2021-11-01
Editors
Sean McGregor, Khoa Lam, Daniel Atherton
Applied Taxonomies
MIT

MIT Taxonomy Classifications

Machine-Classified
Taxonomy Details

Risk Subdomain

A further 23 subdomains create an accessible and understandable classification of hazards and harms associated with AI
 

6.2. Increased inequality and decline in employment quality

Risk Domain

The Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental harms, and (7) AI system safety, failures & limitations.
 
  1. Socioeconomic & Environmental Harms

Entity

Which, if any, entity is presented as the main cause of the risk
 

Human

Timing

The stage in the AI lifecycle at which the risk is presented as occurring
 

Pre-deployment

Intent

Whether the risk is presented as occurring as an expected or unexpected outcome from pursuing a goal
 

Intentional

Incident Reports

Reports Timeline

Incident Occurrence+5
Exclusive: The $2 Per Hour Workers Who Made ChatGPT Safer
+1
The Hidden Workforce That Helped Filter Violence and Abuse Out of ChatGPT
Loading...
Exclusive: The $2 Per Hour Workers Who Made ChatGPT Safer

Exclusive: The $2 Per Hour Workers Who Made ChatGPT Safer

time.com

Loading...
'That Was Torture;' OpenAI Reportedly Relied on Low-Paid Kenyan Laborers to Sift Through Horrific Content to Make ChatGPT Palatable

'That Was Torture;' OpenAI Reportedly Relied on Low-Paid Kenyan Laborers to Sift Through Horrific Content to Make ChatGPT Palatable

gizmodo.com

Loading...
Open AI underpaid 200 Kenyans to perfect ChatGPT then sacked them

Open AI underpaid 200 Kenyans to perfect ChatGPT then sacked them

qz.com

Loading...
OpenAI Apparently Paid People in the Developing World $2/Hour to Look at the Most Disturbing Content Imaginable

OpenAI Apparently Paid People in the Developing World $2/Hour to Look at the Most Disturbing Content Imaginable

futurism.com

Loading...
Sama Cancelled OpenAI's ChatGPT Contract 8 Months Earlier

Sama Cancelled OpenAI's ChatGPT Contract 8 Months Earlier

theinsaneapp.com

Loading...
Kenyan Workers Paid Less Than $2 Per Hour By OpenAI To Work On ChatGPT, Report Says

Kenyan Workers Paid Less Than $2 Per Hour By OpenAI To Work On ChatGPT, Report Says

afrotech.com

Loading...
The Hidden Workforce That Helped Filter Violence and Abuse Out of ChatGPT

The Hidden Workforce That Helped Filter Violence and Abuse Out of ChatGPT

wsj.com

Loading...
Kenyan moderators claim exploitation by OpenAI, Sama in ChatGPT development

Kenyan moderators claim exploitation by OpenAI, Sama in ChatGPT development

swalanyeti.co.ke

Loading...
Exclusive: The $2 Per Hour Workers Who Made ChatGPT Safer
time.com · 2023

Content warning: this story contains descriptions of sexual abuse

ChatGPT was hailed as one of 2022’s most impressive technological innovations upon its release last November. The powerful artificial intelligence (AI) chatbot can generate t…

Loading...
'That Was Torture;' OpenAI Reportedly Relied on Low-Paid Kenyan Laborers to Sift Through Horrific Content to Make ChatGPT Palatable
gizmodo.com · 2023

The seemingly simple, shiny, clean world associated with tech is inevitably almost always simultaneously propped up by something darker hidden just below the surface. From mentally injured content moderators sifting through torrents of vile…

Loading...
Open AI underpaid 200 Kenyans to perfect ChatGPT then sacked them
qz.com · 2023

As San-Francisco-based AI development firm OpenAI’s dream of achieving general artificial intelligence (AI) picks up speed, reports of how it used underpaid Kenyans to remove racism, sexism, and violence from ChatGPT-3's language model are …

Loading...
OpenAI Apparently Paid People in the Developing World $2/Hour to Look at the Most Disturbing Content Imaginable
futurism.com · 2023

Content moderation is grueling, deeply traumatizing job. Workers usually don't last more than a year, and due to the horrible things they have to see, they're often left with lasting PTSD once they leave.

That said, for the internet to func…

Loading...
Sama Cancelled OpenAI's ChatGPT Contract 8 Months Earlier
theinsaneapp.com · 2023

OpenAI, an artificial intelligence company, is accused of paying less than $2 to Kenyan workers to make its ChatGPT chatbot less toxic.

This report is based upon an investigation by Time Magazine. It claims that ChatGPT’s workers in Kenya w…

Loading...
Kenyan Workers Paid Less Than $2 Per Hour By OpenAI To Work On ChatGPT, Report Says
afrotech.com · 2023

ChatGPT, an artificial intelligence (AI) chatbot, has become a trending topic since it hit the tech scene. 

However, an investigative report has been released that details the alleged exploitation to improve it.

According to TIME, OpenAI — …

Loading...
The Hidden Workforce That Helped Filter Violence and Abuse Out of ChatGPT
wsj.com · 2023

This transcript was prepared by a transcription service. This version may not be in its final form and may be updated.

Kate: Hey, it's Kate. Today our producer Annie Minoff is going to bring you a story. It's about a little known part of th…

Loading...
Kenyan moderators claim exploitation by OpenAI, Sama in ChatGPT development
swalanyeti.co.ke · 2023

A group of Kenyan content moderators, who played a significant role in developing the Artificial Intelligence (AI) chatbot ChatGPT, has taken a stand against what they allege to be exploitation and underpayment by OpenAI, the parent company…

Variants

A "variant" is an AI incident similar to a known case—it has the same causes, harms, and AI system. Instead of listing it separately, we group it under the first reported incident. Unlike other incidents, variants do not need to have been reported outside the AIID. Learn more from the research paper.
Seen something similar?

Similar Incidents

Selected by our editors

Meta AI Smart Glasses Reportedly Routed Intimate Imagery to Reviewers at Kenyan Contractor Sama Before Meta Ended Contract

Feb 2026 · 16 reports
Loading...
Facebook Content Moderators Demand Better Working Conditions Due to Allegedly Inadequate AI Content Moderation

Facebook Content Moderators Demand Better Working Conditions Due to Allegedly Inadequate AI Content Moderation

Apr 2020 · 1 report
By textual similarity

Did our AI mess up? Flag the unrelated incidents

Loading...
Microsoft's TayBot Allegedly Posts Racist, Sexist, and Anti-Semitic Content to Twitter

Microsoft's TayBot Allegedly Posts Racist, Sexist, and Anti-Semitic Content to Twitter

Mar 2016 · 28 reports
Loading...
Biased Sentiment Analysis

Biased Sentiment Analysis

Oct 2017 · 6 reports
Previous IncidentNext Incident

Similar Incidents

Selected by our editors

Meta AI Smart Glasses Reportedly Routed Intimate Imagery to Reviewers at Kenyan Contractor Sama Before Meta Ended Contract

Feb 2026 · 16 reports
Loading...
Facebook Content Moderators Demand Better Working Conditions Due to Allegedly Inadequate AI Content Moderation

Facebook Content Moderators Demand Better Working Conditions Due to Allegedly Inadequate AI Content Moderation

Apr 2020 · 1 report
By textual similarity

Did our AI mess up? Flag the unrelated incidents

Loading...
Microsoft's TayBot Allegedly Posts Racist, Sexist, and Anti-Semitic Content to Twitter

Microsoft's TayBot Allegedly Posts Racist, Sexist, and Anti-Semitic Content to Twitter

Mar 2016 · 28 reports
Loading...
Biased Sentiment Analysis

Biased Sentiment Analysis

Oct 2017 · 6 reports

Research

  • Defining an “AI Incident”
  • Defining an “AI Incident Response”
  • Database Roadmap
  • Related Work
  • Download Complete Database

Project and Community

  • About
  • Contact and Follow
  • Apps and Summaries
  • Editor’s Guide

Incidents

  • All Incidents in List Form
  • Flagged Incidents
  • Submission Queue
  • Classifications View
  • Taxonomies

2026 - AI Incident Database

  • Terms of use
  • Privacy Policy
  • b74f812