AIPolicyTracker
Menu

MIT AI Risk Repository · Risk Sub-Category · 47.02.08

Bias and discrimination (bias in training datasets)

Category: Ethical and social risks

Open the MIT Risk Navigator All entries from G'sell2024

Description

"AI experts consider training data to be the most salient source of bias in generative AI models. For example, GPT- 2’s training data comes from outbound links from Reddit, a social network often criticized for hosting anti-feminist content.351 As a result, AI models trained on such data are more likely to produce outputs that reflect these biases."

From Regulating under Uncertainty: Governance Options for Generative AI (G'sell2024), as extracted by the MIT AI Risk Repository (CC BY 4.0).

Classification

Causal entity
Other

Subdomain definition: Unequal treatment of individuals or groups by AI, often based on race, gender, or other sensitive characteristics, resulting in unfair outcomes and representation of those groups.

Real-world incidents in this subdomain

Browse all incidents in this subdomain

How other frameworks describe this risk

  • Discrimination

    A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms (Abercrombie2024)

  • Harms of Representation and Other Biases

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

  • Risks from bias and underrepresentation

    International Scientific Report on the Safety of Advanced AI (Bengio2024)

  • Bias

    International AI Safety Report 2025 (Bengio2025)

  • Bias

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

  • Toxicity and Bias Tendencies

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

  • Biased Training Data

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

  • Broken systems

    Navigating the Landscape of AI Ethics and Responsibility (Cunha2023)

Other entries from G'sell2024