The EU AI Act Explained

ChatGPT to be banned? The EU AI Act, proposed in April 2021, represents a significant milestone in AI governance, setting a precedent for global regulation and reflecting the EU's commitment to ensuring the responsible development and use of AI. Central to the Act is the classification of AI systems based on their potential risks to human safety, fundamental rights, and societal well-being. These classifications are divided into four categories, ranging from minimal risk to high-risk systems, with each category attracting differing levels of regulatory scrutiny.

In this video, we aim to analyze the potential implications of the EU AI Act on ChatGPT, a widely used AI language model developed by OpenAI. ChatGPT has demonstrated remarkable capabilities in generating human-like text responses, making it a valuable tool in various domains, such as customer support, content generation, and personal assistance. However, the Act's risk-based approach prompts us to critically examine how ChatGPT falls with its legislation.

Support the Channel:

Socials:

Sources:

[yiannchrst]

I think I suggested making this video and now it's a reality! I am so happy, even if you hadn't read my original comment.

[caterpillar]

I hope the Transparency Act does not lead to forced licensing. Each artist must have the right to opt out of the system.

[albevanhanoy]

While I am favorable for these regulations on paper, I am highly skeptical regarding enforceability.

[andrasadam]

Sam Altman: AI poses human extinction risk, we need regulation
*EU regulates AI, including his company*
Sam Altman: wait no not like that

[novvak]

Great video. All of this seems very sensible to me. Some AI should not be allowed and a lot of it should be regulated. I might be wrong, but we've heard this "We will leave Europe" too many times. I remember when Facebook threatened to leave, multiple times. Or when GDPR was discussed, there were a lot of companies threatening to leave.
Awesome to learn about this new legislation, in the future we might be really thankful for it :))

[HorrorSFManiac]

I worry this might be a double edged sword. A few weeks ago Google released their AI in 180 countries but not in the EU, the desire to protect citizens is good but if regulations take it too far Europe risks being left behind.

[CalinLaurentiuIlie]

As presented in the video, it seems like a sane approach. However, i am skeptical of the positive impact the regulations will have on education and innovation. And moreover, what is the EUs obsession with being leaders in regulation?

[silveryuno]

To me, the 4 levels of AI regulation seem to be enought for now...
Regulation needs to exist and fast! We already live in world here an AI can write a script for you, an AI can copy your voice read said script and an AI "create" art to go with those two!

[ngocgiabaophan]

I dont know if i should say this or not but the Chinese map in the video is just plain wrong, taiwan is not currently under the PRC’s control. Just want to point out a small mistake, other than this, the video seems informative and well-researched.

[PuNicAdbo]

In my opinion it is a very good thing to do so. Let the americans figure it out the hard way while we in europe take a safe approche and earn the fruits on the developments.

[jonC]

Can you do a video on what will happen with the spanish eu presidency since tvey had to call a snap election that will happen on july 23 having spain taken the presidency weeks before and being expected a change in goverment?

[brianthomson]

Once again the EU mistakenly believes it can 'solve' media-generated problems with a one size fits all, top down set of regulations that can be imposed upon a defined geographical area. People who write AI/ML type applications are not geographically bound and can simply shift their marketing plans to focus on the USA etc. If ANY and ALL medical software that uses AI/ML becomes class III and therefore subject to expensive and bureaucratic delays and regulations (like the device regulations EU2017/745) it will simply not be worth developing such software in the EU and the products will be aimed at the UK&USA market instead. Or not be developed at all.

[iamseamonkey]

i think this is good and strong regulation. the four categories seem sensible and it's being done early enough that it will have a lasting impact. i just hope there is proper enforcement.

[silveryuno]

Americans (and others) are always saying the EU is to slow to do anything. Yet, with american politics specifically, I see them constantly arguing about the same things over and over and over again without dealing with whatever new thing is changing the way the world and people's lifes work.

[alucky]

For me, it seems pretty good, but I mean, I don't know about how it'll be enforce and I even haven't read the law itself, so I can't tell if it is good, bad or need changes 100%

[Ea-pbtu]

The biggest problem I see is that this 1. Will in practice make it impossible for open source AIs to develop and 2. Make it much harder for newcomers to get in as only the big companies will be able to deal with the regulations. So either we’ll once again lose out to US companies and all EU startups will (like they already do) move to the us. The US isn’t significantly richer then us because of luck it’s because their system is better at creating economic value.

[azazeln]

Thank the EU for regulating things that the rest of the world gets imposed without their consent

[eranb]

I definitely agree about risk levels 1, 2 & 4, but I'm not sure about level 3.
ChatGPT should be level 2, and should not be banned or forced to leave. However I don't trust it so much. I've seen a post on Twitter by some professor that ChatGPT gave a wrong answer about his views with fake references.

[edgardebruin]

AI Chat, art or anything else is the future

[JustMe-orxz]

I think it makes sense and its good.And about those threats with leaving EU, i dont think that they will renounce to 500 millions of people market.