MIT AI Risk Repository

Browse AI risks

Name: MIT AI Risk Repository: AI Risk Database (browseable copy)
Creator: MIT AI Risk Initiative
License: https://creativecommons.org/licenses/by/4.0/

2,500 risk entries extracted from 74 frameworks, coded by domain, subdomain, causal entity, intent and timing. Filter, then export the current selection with its licence and citation attached.

Export CSV Export JSON

2,500 entries · page 46 of 50

63.07.01a · Additional evidence

Destabilising Dynamics

Feedback Loops

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.07.05a · Additional evidence

Destabilising Dynamics

Distributional Shift

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.08.00a · Additional evidence

Commitment and Trust

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.08.00b · Additional evidence

Commitment and Trust

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.08.02a · Additional evidence

Commitment and Trust

Threats and Extortion

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.08.03a · Additional evidence

Commitment and Trust

Rigidity and Mistaken Commitments

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.08.03b · Additional evidence

Commitment and Trust

Rigidity and Mistaken Commitments

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.09.00a · Additional evidence

Emergent Agency

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.09.00b · Additional evidence

Emergent Agency

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.09.01a · Additional evidence

Emergent Agency

Emergent Capabilities

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.10.00a · Additional evidence

Multi-Agent Security

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.10.00b · Additional evidence

Multi-Agent Security

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.10.01a · Additional evidence

Multi-Agent Security

Swarm Attacks

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
63.10.01b · Additional evidence

Multi-Agent Security

Swarm Attacks

—

From Multi-Agent Risks from Advanced AI (Hammond2025)
64.01.01a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depiction of human likeness)

Impersonation

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.01.02a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depiction of human likeness)

Appropriated Likeness

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.01.03a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depiction of human likeness)

Sockpuppeting

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.01.04a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depiction of human likeness)

Non-consensual intimate imagery (NCII)

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.01.05a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depiction of human likeness)

Child sexual abuse material (CSAM)

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.02.01a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depictions of non-humans)

Falisification

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.02.02a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depictions of non-humans)

Intellectual Property (IP) Infringement

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.02.03a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Realistic depictions of non-humans)

Counterfeit

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.03.01 · Risk Sub-Category

Misuse tactics that exploit GenAI capabilities (Use of generated content)

Scaling and Amplification

"Automate, amplify, or scale workflows"

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)

Human Intentional Post-deployment
64.03.01a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Use of generated content)

Scaling and Amplification

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
64.03.02a · Additional evidence

Misuse tactics that exploit GenAI capabilities (Use of generated content)

Targeting & Personalisation

—

From Generative AI Misuse: A Taxonomy of Tactics and Insights from Real-World Data (Marchal2024)
65.01.00 · Risk Category

Training Data Risks (Transparency)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.01.01a · Additional evidence

Training Data Risks (Transparency)

Lack of training data transparency

—

From AI Risk Atlas (IBM2025)
65.01.02a · Additional evidence

Training Data Risks (Transparency)

Uncertain data provenance

—

From AI Risk Atlas (IBM2025)
65.02.00 · Risk Category

Training Data Risks (Data laws)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.02.01a · Additional evidence

Training Data Risks (Data laws)

Data usage restrictions

—

From AI Risk Atlas (IBM2025)
65.02.02a · Additional evidence

Training Data Risks (Data laws)

Data acquisition restrictions

—

From AI Risk Atlas (IBM2025)
65.02.03a · Additional evidence

Training Data Risks (Data laws)

Data transfer restrictions

—

From AI Risk Atlas (IBM2025)
65.03.00 · Risk Category

Training Data Risks (Privacy)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.03.01a · Additional evidence

Training Data Risks (Privacy)

Personal information in data

—

From AI Risk Atlas (IBM2025)
65.03.02 · Risk Sub-Category

Training Data Risks (Privacy)

Data privacy rights alignment

"Existing laws could include providing data subject rights such as opt-out, right to access, and right to be forgotten."

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.03.02a · Additional evidence

Training Data Risks (Privacy)

Data privacy rights alignment

—

From AI Risk Atlas (IBM2025)
65.03.03a · Additional evidence

Training Data Risks (Privacy)

Reidentification

—

From AI Risk Atlas (IBM2025)
65.04.00 · Risk Category

Training Data Risks (Fairness)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.04.01a · Additional evidence

Training Data Risks (Fairness)

Data bias

—

From AI Risk Atlas (IBM2025)
65.05.00 · Risk Category

Training Data Risks (Intellectual property)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.05.01 · Risk Sub-Category

Training Data Risks (Intellectual property)

Data usage rights restrictions

"Terms of service, license compliance, or other IP issues may restrict the ability to use certain data for building models."

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.05.01a · Additional evidence

Training Data Risks (Intellectual property)

Data usage rights restrictions

—

From AI Risk Atlas (IBM2025)
65.05.02a · Additional evidence

Training Data Risks (Intellectual property)

Confidential information in data

—

From AI Risk Atlas (IBM2025)
65.06.00 · Risk Category

Training Data Risks (Accuracy)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.06.01a · Additional evidence

Training Data Risks (Accuracy)

Data contamination

—

From AI Risk Atlas (IBM2025)
65.06.02a · Additional evidence

Training Data Risks (Accuracy)

Unrepresentative data

—

From AI Risk Atlas (IBM2025)
65.07.00 · Risk Category

Training Data Risks (Value alignment)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded
65.07.01a · Additional evidence

Training Data Risks (Value alignment)

Improper retraining

—

From AI Risk Atlas (IBM2025)
65.07.02a · Additional evidence

Training Data Risks (Value alignment)

Improper data curation

—

From AI Risk Atlas (IBM2025)
65.08.00 · Risk Category

Training Data Risks (Robustness)

-

—

From AI Risk Atlas (IBM2025)

Not coded Not coded Not coded

Disclaimer: informational only, not legal advice. Verify every claim against the linked official sources and consult a qualified lawyer before acting.

Frequently asked questions

Where do these risk entries come from?

The MIT AI Risk Repository, which extracts risks from dozens of published frameworks, taxonomies and papers and codes each one by domain and subdomain, and by a causal taxonomy of entity, intent and timing. This is a browseable copy, attributed and openly licensed, not original research.

What do entity, intent and timing mean?

They are the causal coding. Entity is whether a human or the AI system is the cause; intent is whether the harm was intentional or not; timing is whether it arises before or after deployment. Together they let you separate misuse from malfunction.

Can I export the results?

Yes, any filtered set exports as CSV or JSON, and every export carries the upstream source, licence and citation with it, because attribution is a condition of the licence rather than a courtesy.

What is the MIT AI Risk Repository?

A living database of AI risks extracted from published frameworks, taxonomies and papers, classified by a causal taxonomy of entity, intent and timing and by a domain taxonomy of seven domains and 24 subdomains. It is published by the MIT AI Risk Initiative under CC BY 4.0.