Is GPT-3 Unsafe?

2 pointsumermirzapk4 comments

I think as humans we adapt. Cars were initially unsafe, we added rules/regulations/technologies to make them safer.

I think GPT-3 and AI in general will follow a similar path. Collectively, we all benefit from these types of technologies. By banning it, we simply build black markets. Instead by investing in AI Safety Ethics and technology, we can harness the benefits and limit the downsides of GPT-3.

True. AI Ethics is a very important concern here. I wonder some big tech giants, collectively or some governing body should be formed that must play a role to enforce regulations around the ethical use of AI.

What do you mean by "unsafe"? Some examples:

Is someone reading all the text you send to the server?

Is is possible that GPT-3 decides to use swears words in the middle of a sentence?

Can GPT-3 output a copy of copyrighted material?

In the context of example #2

Reply on news.ycombinator.com