These visualizations are experimental and still being refined. Designs and data presentation may change.
Where Datasets Disagree
Each subdomain is ranked across four measures (risks, incidents, governance, expert concern). The Δ badge shows how much rankings disagree: a high Δ means datasets treat this risk very differently.
Risk
Incident
Governance
Expert Concern
Subdomain
Δ
24181261
1 Discrimination & Toxicity
Δ10
1.1 Discrimination
Δ5
1.2 Toxic content
Δ16
1.3 Unequal performance
Δ14
2 Privacy & Security
Δ8
2.1 Loss of privacy
Δ8
2.2 AI security vulnerabilities
Δ8
3 Misinformation
Δ10
3.1 False information
Δ13
3.2 Loss of consensus reality
Δ18
4 Malicious Actors & Misuse
Δ3
4.1 Disinformation & influence
Δ4
4.2 AI weapons & cyberattacks
Δ10
4.3 AI fraud & scams
Δ10
5 Human-Computer Interaction
Δ8
5.1 Overreliance & unsafe use
Δ4
5.2 Loss of human agency
Δ12
6 Socioeconomic & Environmental
Δ1
6.1 Power centralization
Δ19
6.2 Inequality & unemployment
Δ6
6.3 Devaluation of human creativity
Δ8
6.4 Competitive dynamics
Δ15
6.5 Governance failure
Δ18
6.6 Environmental harm
Δ10
7 AI System Safety, Failures & Limitations
Δ6
7.1 AI misalignment
Δ16
7.2 Dangerous capabilities
Δ17
7.3 Capability & robustness
Δ19
7.4 Transparency & interpretability
Δ15
7.6 Multi-agent risks
Δ7
Rank 1 = highest value for that metric. The Δ badge shows the difference between the highest and lowest rank across the four metrics. Click any row to view its subdomain detail page.