ChatGPT might quit the EU rather than comply with regulations

AppleInsider

The CEO of ChatGPT developer OpenAI says the firm may pull out of the Europe Union if current draft legislation is not toned down.

ChatGPT

ChatGPT seems to have become instantly omnipresent in 2023, but it may not stick around in Europe. According to Reuters, CEO Sam Altman says the company does not plan to leave, but may need to.

"The current draft of the EU AI Act would be over-regulating, but we have heard it's going to get pulled back," Altman told Reuters. "They are still talking about it."

An EU-wide AI legislation has been in development for years. In 2020, representatives from Apple, Google, and Facebook lobbied the EU over its AI regulation plans.

Speaking about the latest proposals at an industry event in London, Alman said that OpenAI would try to comply with regulation if it can. But currently the planned legislation presents greater hurdles to what are called General All Purpose AI Systems -- such as ChatGPT.

"There's so much they could do like changing the definition of general purpose AI systems," Altman told Reuters. "There's a lot of things that could be done."

Before any of the current concerns about ChatGPT -- concerns that prompted Apple to ban its use on company devices -- the EU has been ahead of the game in looking to ensure that AI can be trusted.

"On artificial intelligence, trust is a must, not a nice-to-have," said Margrethe Vestager, the commission's digital chief, in 2021. "With these landmark rules, the EU is spearheading the development of new global norms to make sure AI can be trusted."

A ChatGPT app for iPhone has recently been released, and is now available in an increasing number of countries and territories.

Read on AppleInsider

Bidli

I will definitely not miss ChatGPT if it has to leave the EU — I'm really worried about the effects of this technology. On the other hand, I don't understand why you would regulate AI while politicians can utter all the crap they want and people still trust them.

avon b7

All algorithm based tech, AI included, must be highly regulated through legislation. 

It's far better to try and lay foundational controls to guide progress than have a free-for-all and try to clean up the mess later. 

As it is, this is still draft legislation and will see modifications as it progresses. 

mikethemartian

How would the ban be practically implemented? The OpenAI servers don’t need to be located within the EU for residents to access them.

larryjw

When first installed, the ChatGPT app highlights the warning that it may return false information, 
The May 10 Lancet published the following article, along with detailed supplemental addendum showing the interaction with ChatGPT which anyone can then reproduce to verify the results. 
The dangers of using large language models for peer review

ChatGPT responded with totally made up material, sounding quite authoritative. 

Banning ChatGPT would seem a good idea. It would give alternative AI systems which actually can tell the truth an opportunity to be developed -- if that can be done. 

Large Language Models like GPT-4 are inherently lying machines. 

22july2013

There are at least 50 large companies in China building products using unmonitored AI software and exporting them to the EU for profit. Does the EU plan on banning any of them? Eg, DJI.

danox

1. Aside from Germany and Holland, all practical industrial/tech things will originate, be built somewhere else outside the EU.

2. Silicon Valley doesn’t have to worry about anything competitive in tech (AI) coming from Europe because of interference from the EU.

zimmie

larryjw said:

When first installed, the ChatGPT app highlights the warning that it may return false information, 
The May 10 Lancet published the following article, along with detailed supplemental addendum showing the interaction with ChatGPT which anyone can then reproduce to verify the results. 
The dangers of using large language models for peer review

ChatGPT responded with totally made up material, sounding quite authoritative. 

Banning ChatGPT would seem a good idea. It would give alternative AI systems which actually can tell the truth an opportunity to be developed -- if that can be done. 

Large Language Models like GPT-4 are inherently lying machines. 

GPT isn't strictly lying, you just aren't asking it the question you think you're asking. People think they're asking the question they put in the prompt. What GPT actually does is produce something which looks like what an expert would say in response to the prompt. For example, experts cite stuff when they're asked about legal or medical matters, so the response should have citations. If there are no high-quality real citations to include, make some up which look plausible.

This behavior becomes much more obvious when you realize GPT (and Bard, and so on) will "believe" anything you tell it is at the other end of something which looks like a URL. The domain doesn't have to resolve, and the path doesn't have to be valid. An expert would load the page and potentially respond with new information based on what they read there. GPT can't actually load a URL or read, but it can fake a response which looks like what an expert who did all that might say.

People have been doing things like telling Bard that Google discontinued Bard months ago, and it responds with something like "Oh. I'm sorry, according to the link you provided, you're correct. Bard was discontinued in March."

Edit: Put "believe" in quotes. As Larryjw pointed out, the model doesn't actually know anything, so it doesn't actually believe anything. Ultimately my point is people think of GPT as being built to produce correct output, when it's actually built to produce plausibly formatted output.

applebynature

mikethemartian said:

How would the ban be practically implemented? The OpenAI servers don’t need to be located within the EU for residents to access them.

It would be the same as the GDPR. Tech companies don't have to have servers, or even a single office or employee based in the EU to fall under its scope. If you allow EU citizens to use your service within the EU, then you must abide by the rules they have imposed. If you do not wish to abide by those rules, you can ban EU citizens from utilizing your services. Sure, an EU citizen could use a VPN based outside of the EU and supply fake information to your company saying you are not within the EU to gain access to your company's systems, but your company would no longer be held liable in that instance, as you have made a good faith effort on your side to ban them.

rotateleftbyte

Go now. Go now. Go now
Don't you even try telling me
That you don't want me to end this way

From the song 'Go Now' by the Moody Blues.

Go ChatGPT, you won't be missed. Most people don't know who you are and don't care.

danox

Bidli said:

I will definitely not miss ChatGPT if it has to leave the EU — I'm really worried about the effects of this technology. On the other hand, I don't understand why you would regulate AI while politicians can utter all the crap they want and people still trust them.

Some Americans are trying to ban abortion. How do you think that’s going to work in a greater world? If the EU tries to do anything significantly restricting to future AI technology, they will just be left behind. The result will be the same when it comes to birth control technology coming up, the world will move on in that area in the US will be left behind.

larryjw

zimmie said:

larryjw said:

When first installed, the ChatGPT app highlights the warning that it may return false information, 
The May 10 Lancet published the following article, along with detailed supplemental addendum showing the interaction with ChatGPT which anyone can then reproduce to verify the results. 
The dangers of using large language models for peer review

ChatGPT responded with totally made up material, sounding quite authoritative. 

Banning ChatGPT would seem a good idea. It would give alternative AI systems which actually can tell the truth an opportunity to be developed -- if that can be done. 

Large Language Models like GPT-4 are inherently lying machines. 

GPT isn't strictly lying, you just aren't asking it the question you think you're asking. People think they're asking the question they put in the prompt. What GPT actually does is produce something which looks like what an expert would say in response to the prompt. For example, experts cite stuff when they're asked about legal or medical matters, so the response should have citations. If there are no high-quality real citations to include, make some up which look plausible.

This behavior becomes much more obvious when you realize GPT (and Bard, and so on) will believe anything you tell it is at the other end of something which looks like a URL. The domain doesn't have to resolve, and the path doesn't have to be valid. An expert would load the page and potentially respond with new information based on what they read there. GPT can't actually load a URL or read, but it can fake a response which looks like what an expert who did all that might say.

People have been doing things like telling Bard that Google discontinued Bard months ago, and it responds with something like "Oh. I'm sorry, according to the link you provided, you're correct. Bard was discontinued in March."

It isn't strictly lying because lying requires intent and a neuronetwork neither thinks or has intent. It doesn't know anything but it hasn't been designed to know when it doesn't know and just generates in general most likely set of words that follows the previous set of words it's been working with. GPT-4 is a neuronetwork with about 90 layers and about a trillion parameters. (Galileo only needed to estimate 2 parameters determine that the force of gravity causes a falling body to trace a parabola.). 

GPT-4 (or 3 or 3.5) doesn't believe anything. It just constructs an enormous set of word associations (n-grams) assigning conditional probabilities of a word given the previous words and then generates the next word (subject to an arbitrary parameter called the Temperature) to add some randomness so it doesn't always choose the most probable. 

r_mari

People can always use VPNS to reach CatGPT if ChatGPT is taken out of Europe.

larryjw

danox said:

1. Aside from Germany and Holland, all practical industrial/tech things will originate, be built somewhere else outside the EU.

2. Silicon Valley doesn’t have to worry about anything competitive in tech (AI) coming from Europe because of interference from the EU.

The public literature on neuro-nets have been building for 20 years. It's in the public domain. The limitation of training neuro-nets is training data and the power and energy required: megawatts vs 30 watts for the human brain. 

badmonk

The EU wants back doors to encryption, the EU wants consumer privacy, the EU wants open App stores, the EU wants to ban chatGPT, the EU wants open standards, the EU wants a single charging cable, the EU wants to be on the technological cutting edge, the EU wants to regulate and fine large technology companies to death.

Seems like the motivation is always more related to anti-Americanism than clear dispassionate thinking.

chutzpah

badmonk said:

The EU wants back doors to encryption, the EU wants consumer privacy, the EU wants open App stores, the EU wants to ban chatGPT, the EU wants open standards, the EU wants a single charging cable, the EU wants to be on the technological cutting edge, the EU wants to regulate and fine large technology companies to death.

Seems like the motivation is always more related to anti-Americanism than clear dispassionate thinking.

Oh grow up.  The EU wants sensible market regulation for the benefit of its people.  If that's the antithesis to the USA's individualist approach then the USA will just have to lump it if it wants to do business in the EU.  The EU has no obligation to make anything easy for big tech companies from the USA just because nerds on the internet get upset. If the USA wants to retaliate with vindictive counter regulations against EU businesses then so be it, but pettiness wins no one any sympathy.

magman1979

How about this crap do the world a favour and just shutdown while their at it???

Any technology whose foundation is based on theft of intellectual property should be vehemently boycotted, as I am boycotting it.

Please LET the door hit you on the way out of the EU!

danox

Galileo, Newton, and Einstein will not be put back in their box, the EU will lose on this, another example is the United States position with high-speed rail. The rest of the world is going forward with it and the US will have to catch up on the run, if the EU persists, they will have to do the same thing with AI catch up on the run.

Xed

magman1979 said:

How about this crap do the world a favour and just shutdown while their at it???

Any technology whose foundation is based on theft of intellectual property should be vehemently boycotted, as I am boycotting it.

Please LET the door hit you on the way out of the EU!

So are you're boycotting Apple product? Whether you like Apple or not, they hav lost many court cases over patentable technologies that are used in their products.

gadgetfreak-apple

mikethemartian said:

How would the ban be practically implemented? The OpenAI servers don’t need to be located within the EU for residents to access them.

The internet is a series of tubes...

michelb76

I think they will leave in the same way that Facebook has left he EU after they threatened to do so...