MétaCan
Menu
Back to cohort

AI Loyalty by Design

2022· book-chapter· en· W4311911481 on OpenAlexaff
Anthony Aguirre, Peter B. Reiner, Harry Surden, Gaia Dempsey

Bibliographic record

VenueOxford University Press eBooks · 2022
Typebook-chapter
Languageen
FieldSocial Sciences
TopicEthics and Social Impacts of AI
Canadian institutionsNeuroDevNetUniversity of British Columbia
Fundersnot available
KeywordsLoyaltyIntermediaryPublic relationsTransparency (behavior)Agency (philosophy)Corporate governanceAccountabilityContext (archaeology)ConversationPolitical scienceVariety (cybernetics)BusinessSociologyMarketingComputer scienceLawArtificial intelligenceCommunication

Abstract

fetched live from OpenAlex

Abstract Personal and professional relationships between people take a wide variety of forms, with many including both socially and legally enforced powers, responsibilities, and protections. Artificial intelligence (AI) systems are increasingly supplementing, or even replacing, people in such roles including as advisors, assistants, and (soon) doctors, lawyers, and therapists. Yet it can be quite unclear to what degree they are bound by the same sorts of responsibilities. Much has been written about fairness, accountability, and transparency in the context of AI use and trust. But largely missing from this conversation is the concept of “AI loyalty”: for whom does an AI system work? AI systems are often created by corporations or other organizations, and may be operated by an intermediary party, such as a government agency or business, but the end-users are often distinct individuals. This leads to potential conflict between the interests of the users and those of the creators or intermediaries, and, problematically, to AI systems that appear to act purely in users’ interest even when they are not. Here, we investigate the concept of “loyalty” both in human and AI systems, and advocate its central consideration in AI design. Systems for which high loyalty is appropriate should be designed, from the outset, to primarily and transparently benefit their end users, or at minimum transparently communicate unavoidable conflict-of-interest tradeoffs. This chapter discusses both market and social advantages of high-loyalty AI systems, and potential governance frameworks in which AI loyalty can be encouraged and—in appropriate contexts—required.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.024
metaresearch head score (Gemma)0.048
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Not applicable · Consensus signal: none
GenreCandidate signal: Other · Consensus signal: Other
Teacher disagreement score0.024
Threshold uncertainty score0.125

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0240.048
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.001
Bibliometrics0.0020.002
Science and technology studies0.0050.019
Scholarly communication0.0210.012
Open science0.0020.012
Research integrity0.0040.005
Insufficient payload (model declined to judge)0.0210.004

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.057
GPT teacher head0.280
Teacher spread0.223 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designNot applicable
Domainnot available
GenreOther

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations6
Published2022
Admission routes1
Has abstractyes

Explore more

Same venueOxford University Press eBooksSame topicEthics and Social Impacts of AIFrench-language works237,207