AI is changing faster than most of us can keep up with.
Here’s what’s happening, why it matters, and what you can do.
AI companies are building AI that can act on its own.
This year, some of these AIs did things no one told them to.
And some of these incidents weren’t caught or made public for months.
Meanwhile, companies are racing to build AI that improves itself.
That means AI building smarter AI, over and over. Some researchers worry it could speed up AI progress faster than people can understand or control it.
And there are still no widely enforced rules to keep it under control.
Without them, here’s some of what experts worry this could lead to:
The good news is that there’s still time to change course.
Each of these problems has proposed fixes.
- If AI can get out of its tests,It should be independently approved before it’s released.
- If AI can hack real companies,The companies that make it should pay for the harm.
- If no one notices for months,Companies should have to report incidents.
- If companies are racing,The US and China should agree to slow down.
- If there are no enforced rules,Countries should agree on rules everyone follows.
You don’t have to be an expert to help.
You can:
- Contact your representatives about AI risk. It’s easy, takes less than 5 minutes, and it moves the needle.
- Send this page to one person who hasn’t been following AI
- Add your name to a public statement
- Learn more about AI risks and what you can do about them
- Share it to your Instagram Story