ai governance
4Clara (PPE, second year) · Naomi (Theology, fourth year) · Standing Committee (officer)
"If a model can draft a statute, who is responsible when it's wrong — the engineer, the minister, or the model?"
OXFORD UNION · HILARY TERM 2026 · THURSDAY DEBATE
Demo motion (generic, not a real 2026 debate): This House would let AI write the law. Three to four speakers per side, the floor cross-questioned, and the coalition that formed across benches now has somewhere to survive past Friday morning.
13 captures · 4 theme clusters · 6 coalition candidates
PROMPTS · THE FLOOR IS ANSWERING
question bringing
What question are you bringing into the chamber tonight?
what could you offer
What evidence or experience could you offer the floor?
who you want to meet
Who do you want to meet after the debate?
infrastructure you rely on
What institution shaped how you think about tonight's motion?
failure others dont see
What is the strongest argument your side won't make?
THEME CLUSTERS · WHAT THE ROOM IS ACTUALLY ARGUING
ai governance
4Clara (PPE, second year) · Naomi (Theology, fourth year) · Standing Committee (officer)
"If a model can draft a statute, who is responsible when it's wrong — the engineer, the minister, or the model?"
policy
4Rajiv (Law, BCL) · Amelia (MPP, Blavatnik)
"Legitimacy of legislation depends on contestable deliberation. A model collapses contestability into output."
data integrity
3Inès (Computer Science DPhil) · Yusuf (Engineering, second year) · Pablo (Economics, second year)
"Working dataset of statutory text classification errors from a frontier model — happy to share under embargo."
trust networks
2Fionn (History, third year) · Saoirse (English, finalist)
"Comparative case studies from the 1832 Reform Act on technological deference and parliamentary capture."
CROSS-BENCH COALITION CANDIDATES · BY SHARED ARGUMENT
Clara (PPE, second year) ↔ Naomi (Theology, fourth year)
ai governance
Rajiv (Law, BCL) ↔ Amelia (MPP, Blavatnik)
policy
Inès (Computer Science DPhil) ↔ Yusuf (Engineering, second year)
data integrity
Inès (Computer Science DPhil) ↔ Pablo (Economics, second year)
data integrity
Yusuf (Engineering, second year) ↔ Pablo (Economics, second year)
data integrity
Fionn (History, third year) ↔ Saoirse (English, finalist)
trust networks
STANDING COMMITTEE · WORKING-GROUP BRIEF
Standing Committee (officer)
Working-group candidates: cross-faculty UK AI eval coalition (Inès, Yusuf, Amelia, Rajiv). Next Thursday floor briefing slot reserved.
CAPTURE FROM THE FLOOR
The argument you almost made. The cross-bench ally you spotted. The standing committee note. Voice or tap — lands in the cluster and (consent-permitting) the coalition map.
AMBIENT CAPTURE · Oxford Union · Hilary 2026
voice · camera · tap · qr
RECENT CAPTURES
Working-group candidates: cross-faculty UK AI eval coalition (Inès, Yusuf, Amelia, Rajiv). Next Thursday floor briefing slot reserved.
ai governanceCoalition forming around model-drafted-bill audit standard. PPE + Law + CS + MPP overlap is the wedge.
ai governanceIf Inès and Yusuf publish a joint UK eval brief by Easter, the policy side has actual evidence to deliver.
policyCross-examining the prop case: contestability isn't a procedure, it's the *only* legitimacy currency the chamber respects.
policyRhetoric is a constitutional act. What does it mean when statutory rhetoric is generated rather than chosen?
trust networksTheir side won't admit that human-drafted law also fails — comparable error rates would strengthen our case.
data integrityOur side won't say it, but speed of legislation is itself a value — sometimes a model-drafted bill is the only one that ships.
ai governanceOpen-source eval suites. If those break, the legal critique of model outputs collapses with them.
data integrity