Do you think AI should be regulated by governments, or should the market regulate it instead?
When I was reading TechCrunch today, I came across Jensen Huang’s (NVIDIA CEO) view that AI safety should be handled by companies through engineering and market pressure, rather than through new regulation.
In my opinion, the power of consumers and the public is sometimes simply not strong enough to counterbalance large technology companies. That’s why I believe governments sometimes need to step in and establish rules and regulations.
Do you think AI should be regulated by governments or by the market?
I personally think that there should be a hybrid form of regulation. I completely dislike the idea of regulating the development and use of AI and how the EU is approaching it, but at the same time I appreciate the initiative that, for example, protects copyright and the ethical use of AI.
Replies
This might be an unpopular opinion, but I think the biggest risks often come from the market and end-developers rather than the AI providers themselves. Anyone can build a project today, and many don't care about safety rules.
For example, most of the chaotic AI automation we see today came from third-party developers using APIs, not from the AI providers themselves. When something goes wrong, people blame the big company. But if you use an API, it should be your mission to make your app safe, not just the provider's job.
@askim_stucki I agree about third-party risks, so my position: government should regulate uses and outcomes, not research or model internals, because AI companies are the only ones who can do the actual safety engineering.
@alieksiaThat’s true but once something is in the market, people will still misuse it. It’s like selling laundry pods and hoping nobody eats them. You can regulate the product, but the user is always the final risk manager.
People fear AI companies.
But the real mess always comes from the outside. hobby devs, rushed startups, automation cowboys.
A coordinated org causes less chaos than the open market ever will.
@alieksia @askim_stuck Interesting point. But then, what is the solution to that mess?
@alieksia @askim_stucki I disagree your example is mild :) It is like a gun or weapon , sure people can misuse it but there needs to be initial regulation on that by goverment.
@askim_stucki Good point, and I think both things can be true at once. Big AI companies build the base model and set some ground rules. But once you hand out an API, thousands of developers build on top of it, and most of the actual risk shows up in what they build, not in the base model itself.
So maybe it's less "who's more at fault" and more "who's responsible for what."
minimalist phone: reduce your screentime
@askim_stucki It is true. But then we should have rules about who will take responsibility for things when they go wrong. AI will not take responsibility; all the consequences are passed on to people.
And it is way easier to have a court case with a trillion-dollar company rather than some hobby dev. It is financially attractive and also good material for media news.
minimalist phone: reduce your screentime
@meetnaz That's funny, to be honest. Some people who do not have expertise and knowledge about the technology are going to decide on it :D At least, they should have dedicated experts.
There should be a balance, but before regulating something we should be able to understand the implications of regulating or not. It should be a universal effort. EU regulators unfortunately dont understand technology, history has shown.
We should start with best practises and alignment from the key players in the AI space, achieving a consensus there and then set some basic laws that will allow the innovation to thrive responsibly.
minimalist phone: reduce your screentime
@dimneo well, and no institution will restrict equally. So while we (in Europe) have 1000 restrictions, US can have only 2 and grow/progress faster.
@busmark_w_nika That's why I left Europe in August and relocated to the US; it is not only the regulate first, innovate after ( if it is possible) mentality. Anyway, that is another conversation.
minimalist phone: reduce your screentime
@dimneo And is it in overal better in the US?
@busmark_w_nika I guess so, the numbers never lie. US has many more unicorns than the EU. Here, they walk the talk, not the talk is walking.
minimalist phone: reduce your screentime
@d_ferencha I like this definition :D
ProdShort
I’m in favor of regulation when it comes to the legal side (fo example privacy, misuse, accountability, etc). But beyond that, I think we should leave enough freedom for AI to develop and innovate without overregulating it.
minimalist phone: reduce your screentime
@amraniyasser I am on the same page + I would add: Who will take the responsibility when things go wrong.
ProdShort
@busmark_w_nika When it comes to responsibility 😂 :
minimalist phone: reduce your screentime
@amraniyasser we all are responsible but none of us want to admit it :D
Dial
Building agent products day to day, my honest take is that regulating "AI" as one thing doesn't really work, because a chatbot and an autonomous agent that can take actions on your behalf carry completely different risk profiles. I'd rather see rules scoped to what the system is allowed to actually do (can it move money, send emails, touch production systems) than rules scoped to the model itself. That also happens to be the layer where a small team can actually build good defaults - permission scopes, human approval steps, audit logs - versus waiting on either a big lab's safety team or a slow-moving regulator to catch up.
minimalist phone: reduce your screentime
@galdayan How would you build the law and taking responsibilities? Because we passed a lot of work on AI (AI can make mistakes), but it doesn't take responsibility for them.
Dial
@busmark_w_nika Practically, I'd tie it to who held the "approve" button. If a human clicked send/approve on the action, they own the outcome, same as approving a wire transfer today. If the system acted with no human in the loop past a certain risk threshold (moving money, deleting data, contacting customers), the liability should sit with whoever deployed it that way, not the model provider - they picked to skip the checkpoint. That gives builders a concrete incentive: add the approval step or accept the downside yourself.
None! Governments are very inefficient, big companies have other incentives.
I believe in entrepreneurship and the power of the free market. Every problem is an opportunity for a new product!
It is in our hands, entrepreneurs, specially the small ones like the PH community, to come up with ideas for new products and projects that could improve AI safety.
Probably, there will be no magic bullet that solves everything, but a lot of small ideas that will address different aspects.
Dangers will always exist; life is difficult! But private initiative is the best tool to make significant contributions to a better world.
Not governments, not big established companies—you, me, all of us.
minimalist phone: reduce your screentime
@narcismirandes But is very difficult to make other people take responsibility. But yeah, we need to start with our own initiative.
What do you think is the biggest challenge with regulating AI without slowing innovation?
minimalist phone: reduce your screentime
@shivam_kushwaha16 I think the fact that we are limited in developing / researching it (esp. under the EU law) so we can fall behind.
minimalist phone: reduce your screentime
@pranay19 Which requires more money... Does it mean we will pay higher taxes? :D
@busmark_w_nika Higher taxes on AI companies :P They are the ones earning maximum from this .
minimalist phone: reduce your screentime
@pranay19 but they will charge us more :D
I lean hybrid too, the real question is whether users can actually see the harm. Market pressure works when bad outcomes are visible. It breaks down for things like training data or safety testing, where the average user just can't see that layer that's an information gap, not market weakness, and it's exactly what disclosure/audit-style regulation is good at filling.
Where I'd push back on heavy regulation is the R&D process itself that tends to just lock in whoever's already big enough to absorb the compliance cost.
minimalist phone: reduce your screentime
@aruna_polosofttech It seems there will be a lot of work regarding these regulations to work out :)