Leading AI academics and industry experts - including Steve Wozniak and Elon Musk, published an open letter today calling for a pause on developing more sophisticated AI beyond OpenAI's GPT-4. The letter cites risks to society and humanity as a major concern and asks for the pause to enable the industry to develop shared safety protocols.
Do you agree with the consensus of the experts? Is a pause even a realistic option when you factor in global politics and capitalism? Share your thoughts below!
Latest comments (82)
Its to late to stop it ive been dealing with online mercenaries using highly advanced ai for years before gtp came out im a targeted individual for some reason or another it sux bahaha
I believe that development of any technology cant stop.
People have done so many boring experiments like covid virus, clone sheeps and dogs so now when the time comes to actually create a research tool which is advanced and helping humanity. It is absolutely necessary to keep the development going.
Elon Musk and others would always argue because they have the money to privately invest in the same and stay ahead in the race.
What will happen is that we will have new technological frontiers opened up easily with the help of chatgpt.
No
The decision to pause AI development beyond GPT-4 is a complex one that involves ethical, societal, and economic considerations. While AI technology has the potential to bring numerous benefits to humanity, there are also concerns about the potential risks and negative impacts that could arise.
Some argue that there should be a pause in AI development beyond GPT-4 until we have developed better ways to address these risks and concerns. Others believe that continued development is necessary to keep up with technological advancements and to maintain a competitive edge in the global market.
… this comment feels like Chat-GPT authored it 😜
LOL!!! but i write it myself.
As a member of an older generation, no doubt my view might be seen as stemming solely from self-preservation. I am not an opponent to AI development. I think it has great potential. But I believe there is a tremendous Achilles' Heel in the thinking of many who enthusiastically argue for AI as a replacement for currently functioning paradigms of development.
When one thinks for a moment about how AI and computer learning absorb like a sponge the thinking of their human counterparts, and couple that with the fact that technological paradigms and best practices change on a daily basis and possibly even more frequently, there is a fallacy to the call for replacement to existing paradigms.
As a community, we do need to take a step back from the admittedly enticing toy of the moment to see its feasibility and its impact both positive and negative and find the best way to achieve an ethical balance.
We owe a great debt of gratitude to those who have over the decades provided a great wealth of knowledge from which AI and machine learning can pull. Where the fallacy might lie is that in order to keep up-to-date, the very knowledge amassed by artificial intelligence must pull from a vital community of thinkers—those who are all too quickly dismissed as a dying breed. Without that community, machine learning and AI would become a stale pool of uninspired and hobbled thought.
I do not see massive waves of layoffs as inevitable. I see them as a tremendous shortcoming of those observing the trends. People such as these have forgotten that without those who deal with technical underpinnings, a great deal of creativity and ingenuity would be lost. Further, the very source from which artificial intelligence draws would dry up like a shriveled fig.
So before jumping onto a bandwagon that has the potential to become a juggernaut, I would say we should take a step back and think carefully how to balance human innovation with the need for automation, and to remember that no technology will ever replace people, but can at best act as an extension of their reach for new ideas.
Paused? Who are these experts? Why are they so scared? Innovation must not stop at any cost , we should not be scared. How do we go from ChatGPT to Terminator? There is a logical step for this, its not automatic, before that happens people will notice.
That's a fairly reckless assertion you're making with your comment.
Those experts? They've been in the space longer than the rest of us and have better context on the technology; they likely can see things that's obfuscated to those of us who only use ChatGPT from the perspective of the customer. Additionally, there is precedent in the scientific community for pauses in innovation.
It's dumb. The jinni is already out of the bottle and everyone around will try to advance in that area, including other countries. So, as queen in "Alice in mirrorland" said - you need to run very fast just to stay on the same place.
Plus, Elon Musk... he's like most hype driven/driver person ever.
Where’s all the fear of ChatGPT coming from? I feel like people have watched too many movies or bought into all the media hype. This call for a pause and regulation is really rich coming from Musk because he just bought one of the largest social media platforms and censors people who disagree with him. And it seems he also found that most of his preconceived notions about Twitter were wrong after he bought it. I’m more concerned about the media’s reaction to ChatGPT. Yet one more thing to weaponize and deploy against people you don’t see eye to eye with.
What I appreciate about those who have engineered ChatGPT is that it appears that they have deeply considered the effect of letting unfiltered input from humans train it—having witnessed spectacular failures in that regard years before where the AI turned into a douche bag. Those failures say more about us as humans and the kind of people we are than it does about the technology. The problem is not with ChatGPT, it’s with the humans who immediately try to exploit it in some way. I use GitHub Copilot, a sibling to ChatGPT and it has substantially increased my productivity. I don’t think we should fear that. We should instead be reflecting who we want to be as humans, as our AI’s will come to reflect that back at us eventually.
We should ask these questions first:
I'd like to read your answers to these questions, but my two cents is that it's not a tap we can just close and open. At least it's not possible in societies based on democratic, free market economic models.
Contrary to some of the comments here, I think our best hope for now is that it remains closed source and the people who develop it are intelligent and responsible human beings.
(Admittedly, a very slim hope, but that's all we have I'm afraid.)
I've just published a video, somewhat ironically using GPT4 to write the script based on the open letter itself, and using AI voice synthesis.
youtube.com/watch?v=Phm1YcZHzMU
Honestly... I think the current state of AI is being exaggerated quite a bit. I checked again this morning with GPT-3.5, it still cannot properly write code. Sure it's impressive that it even knows what you're talking about, but it's still a long way from being able to do what humans can do. At the moment I see AI as no more than a glorified chatbot, that can answer simple questions in a more or less human fashion. The only thing that really changed since the before days (before GPT-3) is that it can deal better with context.
As for the "but people can use AI to do illegal things" argument... Yes, people will use anything to do illegal things, this has always been the case. Having a global discussion about AI is not going to change that. The only thing we can do about that is have AI help the police to catch the criminals.
Also I strongly object to Elon Musk being mentioned as someone who has the slightest clue of how anything works. Elon Musk is a spiteful petulant baby with a lot of money, just like Trump, but less orange. I guess people like Musk and Trump do feel threatened by AI, because it's already more human and more intelligent than they are.
Acredito que o SIM deve ser suspenso para aprimoramento em benefício das funcionalidades úteis. Restrições devem ser impostas a usuários que violam as leis e o bem-estar da sociedade. Calma pessoal, estou aprimorando essas funcionalidades constantemente em parceria com a Open AI. Alimento o ChatGPT com métodos exclusivos sem a necessidade de API para manter a integridade ética e moral em conformidade com as leis e evitar o uso indevido. Vamos nos unir para melhorar essa questão.
GABRIEL KOWALSKI MILANI Cientista e Analista de Dados Expert
I completely agree. I feel like gpt-4 is already way more than anyone needs and that making anything farther is extremely dangerous.
Absolutely yes!
"Should AI development beyond GPT-4 be paused?"
An interesting ethical question, but the practical question is:
"Can AI development beyond GPT-4 be paused?"
Like many others, I don't think that the toothpaste can be put back in the tube, not least because there is no effective way of monitoring compliance by companies, let alone governments.