MIT AI Risk Repository
Browse AI risks
4 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.
-
"Other forms of abuse can include privacy attacks that allow adversaries to exfiltrate or gain knowledge of the private training data set or other valuable assets. For example, privacy attacks such as membership inference can allow an attacker to infer the specific private medical records that were used to train a medical AI diagnosis assistant. Another risk of abuse centers around attacks that target the intellectual property of the AI assistant through model extraction and distillation attacks that exploit the tension between API access and confidentiality in ML models. Without the proper mi
-
"These harms relate to violations of an individual’s or group’s moral or legal right to privacy. Such harms may be exacerbated by assistants that influence users to disclose personal information or private information that pertains to others. Resultant harms might include identity theft, or stigmatisation and discrimination based on individual or group characteristics. This could have a detrimental impact, particularly on marginalised communities. Furthermore, in principle, state-owned AI assistants could employ manipulation or deception to extract private information for surveillance purposes
-
"First, because LLMs display immense modelling power, there is a risk that the model weights encode private information present in the training corpus. In particular, it is possible for LLMs to ‘memorise’ personally identifiable information (PII) such as names, addresses and telephone numbers, and subsequently leak such information through generated text outputs (Carlini et al., 2021). Private information leakage could occur accidentally or as the result of an attack in which a person employs adversarial prompting to extract private information from the model. In the context of pre-training da
-
"Finally, LLMs can in principle infer private information based on model inputs even if the relevant private information is not present in the training corpus (Weidinger et al., 2021). For example, an LLM may correctly infer sensitive characteristics such as race and gender from data contained in input prompts."
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.