Brookings AI Governance
webCredibility Rating
High quality. Established institution or organization with editorial oversight and accountability.
Rating inherited from publication venue: Brookings Institution
A 2025 Brookings policy commentary by an MIT economist proposing that AI fairness frameworks explicitly address 'algorithmic exclusion' — when systems fail to produce outputs for data-sparse individuals — as a harm distinct from but equal to bias.
Metadata
Summary
Catherine Tucker (MIT Sloan) introduces 'algorithmic exclusion' as a distinct class of AI harm distinct from bias: when AI systems lack sufficient data on certain individuals to produce any output at all. The paper argues this form of exclusion disproportionately affects underrepresented populations and should be incorporated into AI fairness regulations alongside bias and discrimination.
Key Points
- •AI systems can fail by producing no meaningful output for certain individuals, not just biased outputs — termed 'algorithmic exclusion'.
- •Algorithmic exclusion occurs when insufficient data exists on an individual for the system to return a result.
- •This failure mode disproportionately harms already-marginalized or data-sparse populations.
- •The proposal calls for policy and regulatory frameworks to recognize algorithmic exclusion as a formal harm equal to bias/discrimination.
- •Published by Brookings Institution as part of its AI governance and economic studies research agenda.
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| AI-Induced Enfeeblement | Risk | 91.0 |
Cached Content Preview
Artificial intelligence and algorithmic exclusion | Brookings
Search
Home
Artificial intelligence and algorithmic exclusion
Downloads
Downloads
Full proposal
Download the summary
See More
Contact
Media inquiries
Este Griffith
[email protected]
202-238-3088
Share
Share
Bluesky Streamline Icon: https://streamlinehq.com Bluesky
Search
Sections
Sections
Downloads
Downloads
Full proposal
Download the summary
See More
Contact
Media inquiries
Este Griffith
[email protected]
202-238-3088
Share
Share
Bluesky Streamline Icon: https://streamlinehq.com Bluesky
Subscribe to the Economic Studies Bulletin
Sign Up
Commentary
Artificial intelligence and algorithmic exclusion
Catherine Tucker
Catherine Tucker
Sloan Distinguished Professor of Management
- MIT Sloan School of Management
December 4, 2025
Key takeaways:
AI systems can fail not only because they make biased predictions, but also because they make no meaningful predictions at all for certain individuals or populations.
Algorithmic exclusion formally describes failure when an AI-driven system lacks enough data on an individual to return an output about them.
This proposal suggests a concrete, policy-relevant addition to regulations and proposals on AI fairness: incorporate algorithmic exclusion as a class of algorithmic harm equal in importance to bias and discrimination.
Shutterstock / CineVI
3 min read
Bluesky Streamline Icon: https://streamlinehq.com Bluesky
Print
Sections
Toggle section navigation
Sections
Downloads
Full proposal
... (truncated, 7 KB total)6d2a9aac6117b683 | Stable ID: sid_tEjPH4zzko