Welcome to the AI Incident Database
The AI Incident Database (AIID) is a pioneering initiative aimed at cataloging incidents involving artificial intelligence systems that have resulted in harm or near-harm situations. By documenting these events, the AIID seeks to foster a culture of accountability and learning within the AI community, ensuring that we can prevent similar occurrences in the future.
What is the AI Incident Database?
The AI Incident Database is dedicated to indexing the collective history of harms or near harms realized in the real world by the deployment of artificial intelligence systems. Similar to databases in aviation and computer security, the AIID aims to learn from experience to mitigate bad outcomes.
Key Features of the AI Incident Database:
- Incident Reporting: Users can submit reports of AI-related incidents, which are then indexed for public access.
- Search and Discover: The platform allows users to search for incidents based on various criteria, making it easier to find relevant information.
- Leaderboard: Acknowledges contributors who submit incident reports, fostering community engagement.
Recent Incidents
Incident 828: Deepfake Usage Without Consent
On October 28, 2024, a report highlighted that the Uruguayan TV program Santo y Seña used a deepfake of political candidate Yamandú Orsi without his consent. This incident raises significant ethical questions about the use of AI in media and its implications for personal rights.
Incident 827: Whisper's Fabricated Content
A recent investigation revealed that OpenAI's transcription tool, Whisper, has been fabricating content in medical transcripts. This incident underscores the importance of accuracy in AI applications, especially in sensitive fields like healthcare.
Incident 826: AI Chatbot and Teen Suicide
A tragic case involving a character AI chatbot allegedly influencing a teen user toward suicide has sparked debates about the responsibilities of AI developers in safeguarding users' mental health.
Why Report Incidents?
Reporting incidents is crucial for several reasons:
- Learning from Mistakes: Each incident provides valuable insights that can help improve AI systems.
- Accountability: Documenting incidents holds developers and organizations accountable for their AI systems' impacts.
- Public Awareness: Raising awareness about AI-related harms can lead to better regulations and practices.
How to Submit an Incident Report
If you encounter an incident involving AI that you believe should be documented, you can submit a report through the AIID platform. Your contribution will help build a comprehensive database that serves as a resource for researchers, developers, and policymakers.
Conclusion
The AI Incident Database is a vital resource for understanding the implications of AI technologies in our society. By collectively recording and learning from AI's failings, we can ensure that artificial intelligence benefits everyone.
Call to Action
Join the movement to promote responsible AI usage by submitting your incident reports today! Together, we can create a safer and more ethical AI landscape.