Companies × risks
General-purpose AI developer
13 of the 15 General-purpose AI developers we reviewed acknowledge at least one AI risk. The most common is discrimination, acknowledged by 12.
AI risks these companies acknowledge
- Discrimination: acknowledged by 12 of 15
- Loss of privacy: acknowledged by 10 of 15
- Devaluation of human creativity: acknowledged by 10 of 15
- AI security vulnerabilities: acknowledged by 9 of 15
- Toxic content: acknowledged by 8 of 15
- Weapons & cyberattacks: acknowledged by 8 of 15
- Environmental harm: acknowledged by 8 of 15
- False information: acknowledged by 7 of 15
- Disinformation & influence: acknowledged by 6 of 15
- Fraud & scams: acknowledged by 6 of 15
- Capability & robustness: acknowledged by 6 of 15
- Dangerous capabilities: acknowledged by 4 of 15
- Transparency & interpretability: acknowledged by 4 of 15
- Overreliance & unsafe use: acknowledged by 3 of 15
- Governance failure: acknowledged by 3 of 15
- Unequal performance: acknowledged by 2 of 15
- Loss of human agency: acknowledged by 2 of 15
- Inequality & unemployment: acknowledged by 2 of 15
- AI misalignment: acknowledged by 2 of 15
- Multi-agent risks: acknowledged by 1 of 15
Companies we reviewed
- IBM acknowledges 19 of 24 AI risks
- Alphabet (Google) acknowledges 14 of 24 AI risks
- Microsoft acknowledges 14 of 24 AI risks
- Salesforce acknowledges 13 of 24 AI risks
- Meta acknowledges 11 of 24 AI risks
- Amazon acknowledges 10 of 24 AI risks
- Baidu acknowledges 8 of 24 AI risks
- NVIDIA acknowledges 5 of 24 AI risks
- Alibaba acknowledges 4 of 24 AI risks
- AMD acknowledges 4 of 24 AI risks
- Apple acknowledges 4 of 24 AI risks
- Tencent acknowledges 4 of 24 AI risks
- Samsung acknowledges 3 of 24 AI risks
- China Mobile acknowledges 0 of 24 AI risks
- China Telecom acknowledges 0 of 24 AI risks