AI risk atlas

Explore this atlas to understand some of the risks of working with generative AI, foundation models, and machine learning models.

Risks are categorized with one of these tags:

  • Traditional AI risks (applies to traditional models as well as generative AI)
  • Risks amplified by generative AI (might also apply to traditional models)
  • New risks specifically associated with generative AI

Background image for risks associated with input

Risks associated with input

Training and tuning phase

Fairness Icon representing fairness risks.

Fairness

Data bias
Amplified
Robustess Icon representing robustness risks.

Robustness

Data poisoning
Traditional
Alignment Icon representing alignment risks.

Value alignment

Data curation
Amplified
Data laws Icon representing data laws risks.

Data laws

Data transfer
Traditional
Data usage
Traditional
Data acquisition
Amplified
Intellectual property Icon representing intellectual property risks.

Intellectual property

Data usage rights
Amplified
Transparency Icon representing transparency risks.

Transparency

Data transparency
Amplified
Data provenance
Amplified
Privacy Icon representing privacy risks.

Privacy

Reidentification
Traditional
Informed consent
Traditional

Inference phase

Privacy Icon representing privacy risks.

Privacy

Intellectual property Icon representing intellectual property risks.

Intellectual property

Robustess Icon representing robustness risks.

Robustness

Evasion attack
Amplified
Extraction attack
Amplified
Prompt leaking
Amplified
Multi-category Icon representing multi-category risks.

Multi-category

Prompt priming
Amplified
Jailbreaking
Amplified
Background image for risks associated with output

Risks associated with output

Fairness Icon representing fairness risks.

Fairness

Decision bias
Traditional
Intellectual property Icon representing intellectual property risks.

Intellectual property

Alignment Icon representing alignment risks.

Value alignment

Misuse Icon representing misuse risks.

Misuse

Nonconsensual use
Amplified
Improper usage
Amplified
Harmful code generation Icon representing harmful code generation risks.

Harmful code generation

Privacy Icon representing privacy risks.

Privacy

Explainability Icon representing explainability risks.

Explainability

Background image for non-technical risks

Non-technical risks

Governance Icon representing governance risks.

Governance

Accountability
Amplified
Legal compliance Icon representing legal compliance risks.

Legal compliance

Societal impact Icon representing societal impact risks.

Societal impact

Job loss
Amplified
Human exploitation
Amplified