← All dilemmas
🧠technology
Should AI labs share dangerous-capability test results with rival countries?
Vote to reveal how SplitVote voters split.
YOUR CHOICE
OR
Anonymous voting. No account required. Results update in real time.
Why this dilemma matters
Every new capability quietly removes a previous choice — the question is whether that trade is worth it. Choosing “Yes. Shared risks need shared evidence” prioritises what the system makes easy; choosing “No. That could hand strategic advantages to rivals” gives more weight to what it makes invisible.
Worth asking yourself
- Who benefits from this, and who absorbs the risk?
- Could you reverse the choice if it backfired?
More Technology Dilemmas
- An AI generates a masterpiece painting with no human creative input. Who owns the copyright?
- A self-driving car's brakes fail. It must choose: swerve into a barrier (killing the passenger) or hit a pedestrian who jaywalked.
- Scientists can upload your consciousness to a computer perfectly. Your biological body must die in the process. Is the digital version still you?
- You can permanently delete all social media from existence. The world becomes slower and less connected, but global mental health improves 40%.