Post

This weekend, The AI Doc debuted across the country. We hope it ignites a large public discussion about the future of AI and where to steer it. But if you’re new to this topic, and just saw the film, you might have questions—Here are some answers to 7 of them 🧵

Warning: mild spoilers. If you haven’t watched The AI Doc: Or How I Became an Apocaloptimist yet, bookmark this thread, go watch it, and come back later. We think everyone should watch it.

1. Is all of this just hype? Answer: No. AI has improved rapidly and continues to become more powerful. It can pilot robot bodies, do hours-long programming tasks, win IMO gold medals, generate realistic video, and much more than fits in 280 characters.

2. AI is still bad at many things; how could it threaten humanity within years? No one knows how far AI companies are from dangerously capable AI, but if current trends continue, maybe not far. And if the companies automate more of their R&D, it might advance even faster.

3. What does “AI is grown, not programmed” mean? Modern AI development starts with a neural network, feeds it lots of data, and lets an algorithm automatically nudge it to better-looking behavior. As a result, no one actually understands how they work.

4. Can we trust AI labs to proceed safely? In 2023, OpenAI said “Currently, we don’t have a solution for steering or controlling a potentially superintelligent AI, and preventing it from going rogue.” This is still true for OpenAI, and its competitors.

5. Would more safety testing help? Because no one understands how AI systems work, safety testing is limited to observing their behavior in controlled tests. This is insufficient, especially now that AI systems are becoming powerful enough to often recognize when they're being tested.

6. Can we train AI to care about humans? Not with current methods, which are based on trial-and-error and observed behavior. These methods cannot ensure that AI actually cares about what we want it to care about, rather than only appearing to care in certain situations.

7. What can we do about it? Racing toward superhuman AI, with our current lack of understanding, is extremely dangerous. To survive, we need an international agreement to halt the race.