Values Statement: We believe AI should cause no harm, but enhance the quality of human life, by proactively adopting our AI Governance framework.

Evidence-based Transparent For governance

AI Incidents

TAIM catalog Map Data source & citation Failure Files™

Facebook Provided Offensive Translation for King of Thailand's Birthday Ceremony

Jul 28, 2020

Facebook's Thai-English translation provided during the live broadcast of the King of Thailand’s candle-lighting ceremony on Thai PBS's page...

Thoughts App Allegedly Created Toxic Tweets

Jul 18, 2020

The Thoughts app, which utilizes OpenAI's GPT-3 for tweet generation, is under scrutiny due to allegations of toxic language towards minorit...

Deepfake Images of Australian Teacher Hannah Grundy and 25 Others Created and Circulated Online by Former Friend

Jul 15, 2020

Deepfake images of Australian teacher Hannah Grundy, along with 25 other women, were circulated online using AI deepfake technology. The per...

International Baccalaureate’s 2020 Calculated-Grades Model Reportedly Lowered Students' Scores and College Prospects

Jul 6, 2020

The replacement of International Baccalaureate final exams with a calculated score during the Covid-19 pandemic has sparked debate over fair...

NarxCare’s Risk Score Model Allegedly Lacked Validation and Trained on Data with High Risk of Bias

Jul 1, 2020

The NarxCare overdose risk algorithm, yet to undergo peer-review validation, raises concerns due to its use of sensitive data like doctor vi...

Facebook’s Political Ad Detection Reportedly Showed High and Geographically Uneven Error Rates

Jul 1, 2020

Facebook's political ad detection performance, unveiled by researchers, is reportedly imprecise and exhibits geographical disparities in err...

WeChat Pay's Facial Recognition Security Evaded by Scammers Using Victims’ Social Media Content

Jul 1, 2020

In China, scammers evaded facial recognition security for online financial transactions on WeChat Pay by using victims' social media content...

Reported AI-Assisted Influence Campaigns by GoLaxy Allegedly Targeting Hong Kong and Taiwan Political Discourse

Jun 30, 2020

Recent leaked documents suggest that Chinese firm GoLaxy employed its 'GoPro' AI system in various influence campaigns. In 2020, it purporte...

YouTube's AI Mistakenly Banned Chess Channel over Chess Language Misinterpretation

Jun 28, 2020

YouTube's AI-driven hate speech detection system erroneously flagged chess content, resulting in bans for creators. This occurred due to the...

Facebook's AI Put "Primates" Label on Video Featuring Black Men

Jun 27, 2020

Facebook's AI inappropriately labeled a video with Black men as one about 'primates,' causing offensive prompts to appear for users. This un...

Miami Police Deployed Facial Recognition to Arrest George Floyd Protestor Allegedly without Cause

Jun 24, 2020

The Miami Police's reported use of facial recognition technology during George Floyd protests has sparked debate, as the alleged application...

Image Upscaling Algorithm PULSE Allegedly Produced Facial Images with Caucasian Features More Often

Jun 20, 2020

The image upscaling algorithm, PULSE (Powered by NVIDIA's StyleGAN), has been under scrutiny for its perceived preference towards generating...

Data source

Incident data is from the AI Incident Database (AIID).

When citing the database as a whole, please use:

McGregor, S. (2021) Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database. In Proceedings of the Thirty-Third Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-21). Virtual Conference.

Pre-print on arXiv · Database snapshots & citation guide

We use weekly snapshots of the AIID for stable reference. For the official suggested citation of a specific incident, use the “Cite this incident” link on each incident page.