Governing Artificial Intelligence: Public Preferences and Regulatory Options
Based on a conjoint survey experiment across seven countries, this paper reveals that citizens globally strongly prefer regulating AI with a focus on safety, public oversight, and international cooperation, highlighting a significant misalignment between these public preferences and current dominant regulatory approaches.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine you are the captain of a massive, high-speed spaceship called "Society." The engine room is full of a new, super-powerful fuel called Artificial Intelligence (AI). This fuel can make the ship go faster than ever before, fix broken parts instantly, and even predict where asteroids will hit. But there's a catch: nobody is 100% sure if this fuel might accidentally blow up the engine, steer the ship off a cliff, or let a stranger take the controls.
In the world of science, this is the field of political science and governance. It's the study of how groups of people make rules to live together. When new, powerful technology arrives, leaders have to decide on three big things:
- The Goal: Should we push the engine to go as fast as possible (Innovation), or should we hit the brakes to make sure no one gets hurt (Safety)?
- The Boss: Should the rules be written by the ship's official captains (Government/Public) or by the mechanics who built the engine (Tech Companies/Private)?
- The Map: Should every ship captain make their own rules (National), or should all the ships in the galaxy agree on one set of rules (International)?
For a long time, the people in charge of these ships (politicians and CEOs) have been arguing over these choices. But they often forgot to ask the most important people: the passengers. Do the people actually living on the ship care about speed, or do they just want to make sure they don't get thrown out an airlock? This paper goes out to ask the passengers exactly that.
The Great Passenger Poll: What Do People Actually Want?
Two researchers, Magnus and Jonas, decided to stop guessing and start asking. They didn't just ask a few friends; they organized a massive, global game show involving 14,239 people from seven different countries: Brazil, China, Germany, India, South Africa, the United Kingdom, and the United States. These countries were chosen because they are all different—some are rich, some are developing, some have strict governments, and some are free.
To make the game fair and fun, they used a special trick called a conjoint experiment. Imagine you are at a pizza shop. Instead of just ordering "pepperoni," you have to choose between two weird, mixed-up pizzas.
- Pizza A: Made by the Mayor, uses a secret recipe to keep you safe, and follows rules made by the whole world.
- Pizza B: Made by a fast-food chain, uses a secret recipe to make the pizza taste amazing (but maybe risky), and follows rules made only by your town.
The researchers asked people to pick which pizza they wanted, over and over again, but they kept changing the toppings (the rules) in the background. They did this for three different "flavors" of AI:
- The Workplace: AI helping (or replacing) your job.
- Policing: AI watching for crime.
- Warfare: AI fighting in wars.
The Big Surprise: Safety Wins, Every Time
When the researchers looked at the results, they found something very clear. The passengers were not confused; they had a very strong opinion.
1. Safety is King (or Queen)
Across all seven countries, people overwhelmingly chose the "Safety" pizza over the "Innovation" pizza. They didn't want the engine to go faster if it meant the ship might crash. In fact, the preference for safety was so strong that it was about four times bigger than the difference between who should be the boss or where the rules should come from. If you asked people to choose between a regulation that makes AI safer or one that makes it more innovative, they picked safety.
2. The People Want the Captain in Charge
When it came to who should write the rules, most people preferred the Government (the public authority) over Tech Companies (private self-regulation). They didn't trust the engine mechanics to write their own safety manuals. They wanted an official captain to be in charge.
3. Let's All Agree on One Map
Finally, when asked if rules should be made by just one country or by the whole world, most people preferred International collaboration. They wanted a global agreement, not just a local one.
Who Cares the Most?
The researchers also dug deeper to see why people felt this way. They found that the people who were most scared of AI were the ones who wanted the strictest safety rules.
- If you think AI is risky, unpredictable, or might hurt you personally, you are much more likely to say, "Stop! Let's make sure it's safe first!"
- Interestingly, it didn't matter if you were a "Lefty" or a "Righty" politically. Your political label didn't change your answer. What mattered was how you felt about the risk.
- There was a small difference between the "Elites" (people with high education, high income, and fancy jobs) and everyone else. The elites were a little less worried about safety and a little more okay with letting companies run things, but even they still preferred safety over speed.
The "Oops" Moment: Leaders Are Out of Sync
Here is the twist in the story. The researchers compared what the passengers wanted with what the leaders are actually doing.
- The European Union (EU) is doing a pretty good job. They are focusing on safety, public rules, and international cooperation. This matches what the passengers want.
- The United States is mostly focused on speed (innovation), letting companies write their own rules, and keeping things national. This is the least popular choice among the passengers, even the Americans.
- China is a mix. They focus on safety and government control (which people like), but they prefer national rules over international ones (which people didn't like as much).
The paper suggests that there is a misalignment. The leaders are driving the ship in a direction that the passengers aren't happy with. If the leaders keep pushing for "fast and free" while the passengers are screaming "slow and safe," the passengers might stop trusting the captains, or even try to take the wheel themselves.
The Bottom Line
This study didn't just guess; it measured the feelings of thousands of real people. It suggests that if governments want to make rules that people will actually follow and trust, they need to listen to the passengers. The message from the crowd is loud and clear: "We don't care how fast the ship goes if we aren't safe. We want the captain to be in charge, and we want the whole fleet to agree on the rules."
The researchers admit they only took a snapshot in time (December 2025), and things might change as AI gets better or worse. But for now, the data shows that when it comes to the future of AI, the public is ready to hit the brakes to make sure everyone survives the ride.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.