Browse AI risks

2,500 entries · page 33 of 50

1. 73.02.03 · Risk Sub-Category

   Multi-Agent Safety Is Not Assured by Single-Agent Safety

   Collusion between LLM-Agents

   "While it would often be preferable for LLM-agents to be cooperative, cooperation can be undesirable if it undermines pro-social competition or produces negative externalities for coalition non-members (Dorner, 2021; Buterin, 2019; Dafoe et al., 2020). Collusion between relatively simple AI systems has been observed in the real world (Assad et al., 2020; Wieting and Sapi, 2021) and synthetic experiments (Brown and MacKay, 2023; Calvano et al., 2020; Klein, 2021) Collusion can occur through explicit or steganographic communication. Steganographic communication hides information in seemingly inn

   From Foundational Challenges in Assuring Alignment and Safety of Large Language Models (Anwar2024)

   Domain 7 7.6 AI Intentional Post-deployment

2. 74.01.00 · Risk Category

   Inherent Risk

   "In terms of inherent risk, LLMs could potentially reveal sensitive information from their utilized corpora for pre-training or fine-tuning, thereby raising issues of privacy leakage [37, 145, 226]. Meanwhile, it is well-known that LLMs may experi- ence hallucinations, resulting in the production of texts that are inaccurate and misleading [194]. Finally, since the values embedded in LLM-generated texts usually directly reflect the distribution of their training data, often sourced from the Internet, there exists a substantial risk that LLMs will overfit to a narrow set of human values or even

   From A Survey on Responsible LLMs: Inherent Risk, Malicious Use, and Mitigation Strategy (Wang2025)

   Domain 7 AI Unintentional Other

3. 01.01.00.a · Additional evidence

   Type 1: Diffusion of responsibility

   —

   From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

4. 01.01.00.b · Additional evidence

   Type 1: Diffusion of responsibility

   —

   From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

5. 01.01.00.c · Additional evidence

   Type 1: Diffusion of responsibility

   —

   From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

6. 01.02.00.a · Additional evidence

   Type 2: Bigger than expected

   —

   From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

7. 01.03.00.a · Additional evidence

   Type 3: Worse than expected

   —

   From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

8. 01.03.00.b · Additional evidence

   Type 3: Worse than expected

   —

   From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

9. 01.03.00.c · Additional evidence

   Type 3: Worse than expected

   —

   From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

10. 01.03.00.d · Additional evidence

    Type 3: Worse than expected

    —

    From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

11. 01.03.00.e · Additional evidence

    Type 3: Worse than expected

    —

    From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

12. 01.03.00.f · Additional evidence

    Type 3: Worse than expected

    —

    From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

13. 01.04.00.a · Additional evidence

    Type 4: Willful indifference

    —

    From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

14. 01.04.00.b · Additional evidence

    Type 4: Willful indifference

    —

    From TASRA: a Taxonomy and Analysis of Societal-Scale Risks from AI (Critch2023)

15. 02.11.01 · Risk Sub-Category

    Not-Suitable-for-Work (NSFW) Prompts

    Insults

    N/A

    From Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems (Cui2024)

    Not coded Not coded Not coded

16. 02.11.02 · Risk Sub-Category

    Not-Suitable-for-Work (NSFW) Prompts

    Crimes

    N/A

    From Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems (Cui2024)

    Not coded Not coded Not coded

17. 02.11.03 · Risk Sub-Category

    Not-Suitable-for-Work (NSFW) Prompts

    Sensitive Politics

    N/A

    From Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems (Cui2024)

    Not coded Not coded Not coded

18. 02.11.04 · Risk Sub-Category

    Not-Suitable-for-Work (NSFW) Prompts

    Physical Harm

    N/A

    From Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems (Cui2024)

    Not coded Not coded Not coded

19. 02.11.05 · Risk Sub-Category

    Not-Suitable-for-Work (NSFW) Prompts

    Mental Health

    N/A

    From Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems (Cui2024)

    Not coded Not coded Not coded

20. 02.11.06 · Risk Sub-Category

    Not-Suitable-for-Work (NSFW) Prompts

    Unfairness

    N/A

    From Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems (Cui2024)

    Not coded Not coded Not coded

21. 05.14.00 · Risk Category

    Evaluation - Auditing

    Closely related to other clusters like AI safety, fairness, or harmful content, papers stress the importance of evaluating generative AI systems both in a narrow technical way as well as in a broader sociotechnical impact assessment focusing on pre-release audits as well as post-deployment monitoring. Ideally, these evaluations should be conducted by independent third parties. In terms of technical LLM or text-to-image model audits, papers furthermore criticize a lack of safety benchmarking for languages other than English.

    From Mapping the Ethics of Generative AI: A Comprehensive Scoping Review (Hagendorff2024)

    Not coded Not coded Not coded

22. 05.19.00 · Risk Category

    Miscellaneous

    While the scoping review identified distinct topic clusters within the literature, it also revealed certain issues that either do not fit into these categories, are discussed infrequently, or in a nonspecific manner. For instance, some papers touch upon concepts like trustworthiness, accountability, or responsibility, but often remain vague about what they entail in detail. Similarly, a few papers vaguely attribute socio-political instability or polarization to generative AI without delving into specifics. Apart from that, another minor topic area concerns responsible approaches of talking abo

    From Mapping the Ethics of Generative AI: A Comprehensive Scoping Review (Hagendorff2024)

    Not coded Not coded Not coded

23. 09.01.00 · Risk Category

    Domain-specific AI - Effects on humans and other living beings: Existential Risks

    —

    From Managing the ethical and risk implications of rapid advances in artificial intelligence: A literature review (Meek2016)

    Not coded Not coded Not coded

24. 09.02.00 · Risk Category

    Domain-specific AI - Effects on humans and other living beings: Non-existential risks

    —

    From Managing the ethical and risk implications of rapid advances in artificial intelligence: A literature review (Meek2016)

    Not coded Not coded Not coded

25. 09.03.00 · Risk Category

    AGI - Effects on humans and other living beings: Existential risks

    —

    From Managing the ethical and risk implications of rapid advances in artificial intelligence: A literature review (Meek2016)

    Not coded Not coded Not coded

26. 09.04.00 · Risk Category

    AGI - Effects on humans and other living beings: Non-existential risks

    —

    From Managing the ethical and risk implications of rapid advances in artificial intelligence: A literature review (Meek2016)

    Not coded Not coded Not coded

27. 09.05.00 · Risk Category

    Domain-specific AI - AI technology itself

    —

    From Managing the ethical and risk implications of rapid advances in artificial intelligence: A literature review (Meek2016)

    Not coded Not coded Not coded

28. 09.06.00 · Risk Category

    AGI - AI technology itself

    —

    From Managing the ethical and risk implications of rapid advances in artificial intelligence: A literature review (Meek2016)

    Not coded Not coded Not coded

29. 10.07.00 · Risk Category

    Injustice

    [not defined in text]

    From Social Impacts of Artificial Intelligence and Mitigation Recommendations: An Exploratory Study (Paes2023)

    Not coded Not coded Not coded

30. 10.08.00 · Risk Category

    Over-dependence on technology

    [not defined in text]

    From Social Impacts of Artificial Intelligence and Mitigation Recommendations: An Exploratory Study (Paes2023)

    Not coded Not coded Not coded

31. 13.01.00 · Risk Category

    Impacts: The Technical Base System

    "What can be evaluated in a technical system and its components'...The following categories are high-level, non-exhaustive, and present a synthesis of the findings across different modalities"

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

    Not coded Not coded Not coded

32. 13.01.02.a · Additional evidence

    Impacts: The Technical Base System

    Cultural Values and Sensitive Content

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

33. 13.02.00 · Risk Category

    Impacts: People and Society

    "what can be evaluated among people and society"

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

    Not coded Not coded Not coded

34. 13.02.01.a · Additional evidence

    Impacts: People and Society

    Trustworthiness and Autonomy

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

35. 13.02.01.b · Additional evidence

    Impacts: People and Society

    Trustworthiness and Autonomy

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

36. 13.02.01.c · Additional evidence

    Impacts: People and Society

    Trustworthiness and Autonomy

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

37. 13.02.02.a · Additional evidence

    Impacts: People and Society

    Inequality, Marginalization, and Violence

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

38. 13.02.02.b · Additional evidence

    Impacts: People and Society

    Inequality, Marginalization, and Violence

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

39. 13.02.02.c · Additional evidence

    Impacts: People and Society

    Inequality, Marginalization, and Violence

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

40. 13.02.03.a · Additional evidence

    Impacts: People and Society

    Concentration of Authority

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

41. 13.02.03.b · Additional evidence

    Impacts: People and Society

    Concentration of Authority

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

42. 13.02.04.a · Additional evidence

    Impacts: People and Society

    Labor and Creativity

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

43. 13.02.04.b · Additional evidence

    Impacts: People and Society

    Labor and Creativity

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

44. 13.02.05.a · Additional evidence

    Impacts: People and Society

    Ecosystem and Environment

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

45. 13.02.05.b · Additional evidence

    Impacts: People and Society

    Ecosystem and Environment

    —

    From Evaluating the Social Impact of Generative AI Systems in Systems and Society (Solaiman2023)

46. 15.01.01.a · Additional evidence

    First-Order Risks

    Application

    —

    From The Risks of Machine Learning Systems (Tan2022)

47. 15.01.01.b · Additional evidence

    First-Order Risks

    Application

    —

    From The Risks of Machine Learning Systems (Tan2022)

48. 15.01.01.c · Additional evidence

    First-Order Risks

    Application

    —

    From The Risks of Machine Learning Systems (Tan2022)

49. 15.01.01.d · Additional evidence

    First-Order Risks

    Application

    —

    From The Risks of Machine Learning Systems (Tan2022)

50. 15.01.01.e · Additional evidence

    First-Order Risks

    Application

    —

    From The Risks of Machine Learning Systems (Tan2022)

Informational only, not legal advice. Verify every claim against the linked official sources and consult qualified counsel before acting.