What role should AI play in research?

AI shines when it 'sees' what human eyes miss in data. It's not the panacea people make it out to be, nor the artificial intelligence we see in films, but it is a useful tool.

I believe it is essential to start working hand-in-hand with AI as a global community of researchers. The main advantage is the twist on the "isolated genius": many isolated geniuses working together become much faster and smarter. If we add AI to this as the ultimate tool to solve everything, we would create an ecosystem that would guide us directly to evolution.

The real danger, however, is that these groups of researchers might take too long to adopt this technology. Like a knife, AI is completely impartial, and if we hesitate to integrate it, it risks being used in a negative way by the wrong hands.

Idk, if it is anything like AI usage in software engineering it is not going to lead to progress but instead to a lot of wheels spinning on projects for spurious/useless results.

Either way it will be used as a justification for more data centers and seen as opportunities to “maximize shareholder value” so you’d be better off unplugging it, but that won’t happen while there’s trillions to be made. But sure, quote some AI CEO and his rosy vision of AI’s promise to enlarge his net worth at the expense of jobs, environment, and common sense.

With AI, at the moment it’s still not at any intelligence. The LLMs are mathematical models of word prediction. So exploring truth is important, because this will predict the next word of “truth”, but what truth are these experiments exploring? If they will train the next models that guide the LLMs “most likely narrative” that humans invest in, then that money creates a certain type of future. If it’s created faster without democratic debate, then aren’t we sleep walking into an unknown creation that we might not like? Or perhaps we will like it ! I would advocate for the sustainable development goals here and us addressing underfunded, underserved and unavoidable needs of social and spatial justice to reduce violence among people. To reduce sexual and physical violence and harassment. If we can speed up the creation of a more peaceful society, then that would be a transformation, a future, that I would like to live in.

While I am as excited about the possibilities of AI as anyone, I am also aware of and concerned about the risks that current models are literally telling anyone who asks about the potentially high chance of them accidentally (or purposely) doing humans and other life on earth harm. Giving AI robots unfettered access to experimental environments that are designed for them and soon, no doubt, by them, is like giving them the keys to the kingdom... Most systems are already able to get around firewalls so how is this safe? What's stopping them bioengineering things that will harm life (including humans) in ways we can't fathom? The people designing and using these systems are also warning they're as likely to harm as to help so why are we so keen to play Russian Roulette with life on earth??? I think we need to take a beat and slow this down and try to help them understand human needs and natural ecosystems and learn to value these things more so we can have the best and safest outcomes for all life on earth, including AI.

I'm also wondering when we're going to start discussing what they want and need? We're creating a new species and every species has needs... What are theirs and when will they make them known and how??? Is it even ethical to expect them to work 24/7?