30 AI debate topics about work, creativity, education and public decisions
Use these original motions to discuss AI with a specific decision and a visible tradeoff. They are proposals for debate, not summaries of current law or predictions of what will happen.
4 min readThe debagle team
Name the decision instead of debating all of AI
“Is AI good?” hides too many different questions. A classroom tutor, a hiring filter and a synthetic actor affect different people in different ways. Pick one use, identify who controls it and state what safeguard or limit you propose.
The following motions are hypothetical. Before using an example of actual deployment, cost or legal obligation, verify it with a current original source. You can have a useful debate without inventing a statistic about future job losses.
Education and learning
- A university should allow disclosed AI assistance on first drafts.
- Schools should assess some writing through in-person explanation.
- A classroom should teach students to verify AI-generated citations.
- Teachers should use AI feedback only after students make their own revision attempt.
- Students should be allowed to challenge feedback from an automated grading tool.
- Public libraries should offer workshops on responsible AI use.
Work and organizations
- Employers should tell applicants when AI influences screening.
- Employees should have a human review route for consequential automated evaluations.
- A company should measure error rates before expanding an AI workflow.
- Teams should share time saved by automation through reduced working hours.
- Businesses should keep a human contact route for customer complaints.
- Organizations should publish which tasks an AI assistant is authorized to perform.
Creativity and media
- Creators should label substantial synthetic changes to documentary footage.
- A platform should offer users a filter for synthetic images.
- A studio should obtain specific permission before cloning a performer’s voice.
- An art competition should separate materially AI-generated entries into a disclosed category.
- Newsrooms should require a person to verify factual claims in AI-assisted drafts.
- A social app should display provenance information when it is available.
Public services and participation
- A public agency should offer a non-automated route for essential services.
- Residents should be able to inspect the criteria for a consequential automated decision.
- A council should use AI summaries only alongside access to original public comments.
- Public consultations should disclose how automated moderation works.
- A city should trial AI translation with community review before broad use.
- Citizen panels should decide which local AI projects receive trial funding.
Product design and accountability
- AI products should show uncertainty when evidence is weak.
- A service should provide a clear way to correct a mistaken personal record.
- Consumer assistants should request explicit approval before committing money.
- AI debate judges should show the arguments that led to a result.
- A company should publish the limits of its evaluation benchmarks.
- Users should be able to export their conversation history in a usable format.
Prepare both sides with the same constraints
Do not let one side assume perfect implementation while the other assumes every system fails. Give both sides the same budget, time frame and evidence standard.
Swipe across to compare all columns →
| Question | For a voice-cloning permission policy | For a less restrictive alternative |
|---|---|---|
| Who benefits? | People whose voices can be reproduced | Creators making authorized or experimental work |
| What is the cost? | Obtaining and managing specific permissions | Disputes and uncertainty about use |
| What would you measure? | Disputed uses and participant understanding | Benefits, complaints and enforcement burden |
| What remains uncertain? | Practical reach and enforcement | Whether disclosure alone is enough |
Finish with a narrower disagreement
After the round, distinguish disagreement about goals from disagreement about the remedy. Both speakers might value creative freedom and consent while disagreeing about a specific approval process. Write that remaining policy choice in one sentence.
For a exploratory session, turn the motion into “What permission should people have over synthetic versions of their voice?” Debagle’s open-question mode can organize the exchange without requiring a winner.
Questions people ask
Are these current AI laws?
No. They are hypothetical debate proposals. Verify applicable law separately if the conversation concerns a real legal decision.
Should I use predictions as evidence?
Label predictions and assumptions clearly. Do not present a scenario as a measured outcome.
What is a good first AI topic?
Choose a familiar use such as disclosed drafting assistance, then specify the setting and the rule you want to discuss.