Thoughts: Musk et al’s 6 month proposed moratorium on ‘Big AI’, GPT-4
I think it’s encouraging that AI creators like Elon Musk and, not yet a signatory, OpenAI’s Sam Altman, are independently talking about, and arguably taking seriously, the threats, as well as the benefits, of AI.
So:
- Is the 6 month moratorium a good idea?
- Will OpenAI and others comply?
- What other alternatives exist?
- Is AI really beginning to look like a near-term threat?
- How will this proceed now?
- What does regulation look like?
Let’s take a quick look.
The open letter
It’s important to note that the open letter — of March 29th 2023 — from Future of Life Institute, a futurism think tank, does not propose everyone stops research on their models. It’s just about ‘Big AI’, in fact anything more powerful than the recently released GPT-4.
So that, at face value at least, will only affect a handful of known LLM (large language model) developers like OpenAI, Microsoft, Google, Stability AI and a few startups and open source projects like GPT-J.
The letter makes the appeal:
Contemporary AI systems are now becoming…