I ran across the AI Incident Database incidentally (see what I did there? 😉) last week. Here’s more about it and what it tracks.

The AI Incident Database (AIID) is dedicated to indexing the collective history of harms or near harms realized in the real world by the deployment of AI systems. Like similar databases in aviation and computer security, the AI Incident Database aims to learn from experience so we can prevent or mitigate bad outcomes. As their About page states:

“Intelligent systems are currently prone to unforeseen and often dangerous failures when they are deployed to the real world. Much like the transportation sector before it (e.g., FAA and FARS) and more recently computer systems, intelligent systems require a repository of problems experienced in the real world so that future researchers and developers may mitigate or avoid repeated bad outcomes.”

Current examples include:

• An autonomous car kills a pedestrian
• A trading algorithm causes a market “flash crash” where billions of dollars transfer between parties
• A facial recognition system causes an innocent person to be arrested

They encourage you “to submit incident reports, whereupon submissions will be indexed and made discoverable to the world.”

The site goes back as far as 2015 and has tracked at least 1,500 incidents since its inception. Although the Search box at the top of the site says you can “search over 6,000 reports of AI harms”. I’m not sure what the rest of them are.  You can click on “Discover Incidents” in the left-side menu and change the “Display Option” to “Incident and Issue Reports”, which gives you 6,330 results found – which makes it seem that the majority of the results are Issue Reports, but when you change the “Display Option” to “Issue Reports”, you only get 403 results found. Hmmm.

Regardless, you can get an “Incident List”, which is a simple numeric listing of all incidents and their reports within the database (sorted in descending order by default), or you can get the same list in Table View. Here’s a sample of some of the most recent incidents listed:

• Grok Reportedly Generated Images After X Users Asked It to Unredact Epstein-File Photos of Children and Women
• Meta AI Support Bot Reportedly Enabled Takeovers of High-Profile Instagram Accounts
• ChatGPT Was Alleged to Have Reinforced Pittsburgh Man’s Stalking and Threats Against Women
• Scammers Reportedly Used AI-Cloned Daughter’s Voice to Defraud Bay Area Mother in Fake Kidnapping Call

The site even has a blog and (perhaps most interestingly) an AI News Digest page, with news stories from around the web matching a set of AI-related keywords. Oh, and it has a Leaderboard of the top submitters, authors, and domains by count. Some guy named Daniel Atherton has 825 submissions – over half of the incidents – with the next closest (Anonymous) at 157. What an overachiever! 😉 Seriously though, he’s the lead editor with the AI Incident Database, so it makes sense he would have a lot of them.

Still getting a sense of the information contained in the database, but it seems like a potentially useful resource – at least for me when I’m looking for AI gone wrong stories for the Friday Kitchen Sink. 🤣

So, what do you think? Are you interested in checking out the AI Incident Database? Please share any comments you might have or if you’d like to know more about a particular topic.

Image created using DALL-E 3, using the term “robot lawyer wearing a suit in a law office slipping on a small banana peel”.

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.