I get the argument, but the absolute dichotomy is misplaced, if not a bit disingenuous.
There's nothing stopping regulation of closed-source commercial AI, while promoting academic and opensource use conforming to ethical guidelines.
There's nothing stopping regulation of closed-source commercial AI, while promoting academic and opensource use conforming to ethical guidelines.
These two things are orthogonal, and need to be considered seriously and separately: how to ensure that LLMs are not enabling serious abuses, through various kinds of regulation, and also how to ensure that our common usage of LLMs is not gatekept by huge corporations and dependent on their goodwill and continued interest in supporting the particular things we want to do with them.
All this brings me back to the Vulnerable World Hypothesis proposed by Nick Bostrom. In fact global totalitarian surveillance that stamps out all potentially dangerous research is one of his few proposed solutions in his paper. I don't know if I can stomach such a solution, but I think we are living the early stages of such a world that forces such questions to be asked, not one in which we picked out a "black ball" that will make us go extinct, but a sort of "green ball" that irreversibly poisons specific parts of collective humanity.
I would personally be in favor of OpenAI keeping GPT-4 and future models proprietary, much as it won't affect the open source spirit. Random hackers with powerful GPUs generating terabytes of SEO spam are less visible to the world and more difficult to hold publicly accountable than a huge corporation with billions in funding.
2) There's a big difference between commercial and academic research, particularly in terms of incentives (though the gap has narrowed significantly in recent years). But, research is rarely "disney villain" material to be stamped anyway; good research can be used for evil and vice versa. But at least in academia you're supposed to have ethics committees and academic standards to uphold.
3) I dont accept this argument. It has a very "ban encryption / think of the children" vibe to me. The idea that openness makes the life of spammers ever-so-slightly easier and therefore we should not have openness is bizzare. Spammers will spam regardless, and spamming tools being proprietary isn't exactly a deterrant. The only thing you lose by blocking openness is ... well, openness.