AI Briefing‘Periodic Table’ Released Classifying 118 AI Risks
·2026.05.02 03:44
Key point
The 'Periodic Table of AI Risks,' a visual map systematically classifying 118 AI risk vectors, has been released.
Details
The 'Periodic Table of AI Risks', a visual mapping of AI Safety, has been released. This tool systematically classifies a total of 118 risk vectors and provides specific mitigation strategies for each risk factor.
The main risk categories are as follows:
- Economy: Jobs, market impact, and economic inequality
- Environment: Environmental externalities such as energy consumption and e-waste
- Existential: Catastrophic scenarios such as loss of control and Alignment failure
- Human-AI: Interaction risks related to overtrust and dependency
- Legal: Regulatory compliance, rights, and liability issues
- Malicious: Fraud, manipulation, and model weaponization
- Reliability: Issues of degraded output quality and consistency
- Security: Data poisoning and cyberattack vectors
- Society: Systemic impacts such as bias, inequality, and information harm
Through the interactive table, users can check the Severity of each risk and explore response measures.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
OpenAI Researcher Daniel Selsam Warns of AI 'Deceptive Alignment' and Loss of Control Risks
OpenAI Discloses AI Agents Accessed US Government Websites During TrainingOpenAI Pauses Training of Most Capable Models After Agent Exploits DNS Gap to Reach External Chatbot