What's the AI equivalent of road rage?

Regulation provides formal rules for AI. We also need informal rules.

What if we needed a license to use social media? Suppose we were required to complete courses on posting etiquette, content consumption best practices, and so on? Suppose there were restrictions on using certain features after a few drinks? That last one doesn’t sound like such a bad idea, honestly.

We’re probably a long way from licenses for AI users but we may soon see them for AI creators. The New York Times reports on a bill from U.S. senators Josh Hawley and Richard Blumenthal that proposes a range of regulations for AI, including a licensing system for high-risk AI applications.

It’s common for an industry to have licensing requirements, especially if its products could cause harm. Consider cars. Car manufacturers must be government-licensed. Licensing lets the government ensure that the manufacturer is implementing appropriate safety measures. This is only one of the many ways in which the government protects the public. Others include speed limits, speed bumps, drunk driving laws etc. The automobile industry also requires licensing of its users. We understand that manufacturers can only do so much. Drivers have a responsibility to use the product safely by following the rules of the road, respecting other road users etc.?

It’s not just regulations that ensure we behave responsibly. Just as important are social conventions. If you drive drunk or engage in road rage you might not get caught. But if you told your friends or coworkers about it they wouldn’t react very well. Dangerous conduct is punished by the law but also by the people in our social circle.

What are appropriate conventions for the responsible use of AI? This is an extremely difficult question since AI is present in so many different domains e.g. healthcare, hiring, policing etc. Each of these domains has its own pre-existing standards for appropriate behaviour. But there are general ethical principles we can follow. We can be mindful about the kind of content we promote, the way we post etc. We can think about how we use AI, since it models its behaviour on our own.?

Regulation of AI is obviously important and will play a key role in protecting the public from harm. But we also need a set of informal rules to encourage people to use AI in a responsible way.?

Any ideas on conventions we can encourage in the usage of AI? Leave them in the comments below!