EU lawmakers vote for tougher AI rules as draft moves to final stage

  • last year

Visit our website:
http://www.france24.com

Like us on Facebook:
https://www.facebook.com/FRANCE24.English

Follow us on Twitter:
https://twitter.com/France24_en
Transcript
00:00 European MPs have voted massively in favour of the world's first sweeping rules on artificial
00:06 intelligence systems. The aim then to curb potential harms while nurturing innovation
00:11 at the same time. Although the EU's plan does date back to 2021, the draft rules took on
00:16 greater urgency when ChatGPT exploded onto the scene last year. There is of course a
00:23 growing urgency to regulate AI across the Atlantic as pressure grows on Western governments
00:29 to act swiftly. Let's get you more on today's vote. France 24's Armin Georgian is standing
00:35 by for us. Armin, in terms of this vote, walk us through what happened. Apparently they've
00:40 voted massively in favour for these regulations.
00:44 Yeah, so this is the European Parliament giving its position on the Artificial Intelligence
00:52 Act. So the Parliament has agreed to move forward with this text, but of course it's
00:58 still has to be negotiated in what's called a trilogue with the European Commission and
01:03 with EU member states. So this is by no means the end of the process. However, Rochelle,
01:08 MEPs like to point out that this is the first time that any legislative body in the world
01:14 has tried to regulate artificial intelligence. It's a text that went through quite a few
01:20 negotiations and changes. So a consensus was reached and this vote was passed overwhelmingly
01:26 in favour. 499 MEPs voting in favour of this text, only 28 against and 93 abstentions.
01:36 All right, so that's a big majority, as you say, voting and saying, look, you know, we'd
01:41 like to see some rules put in place with this. It's a fine balance though, isn't it, Armin?
01:45 Because essentially the aim is to kind of curb the riskier parts of AI, but ultimately
01:51 make sure that innovation is respected. Yeah, I mean, the line that MEPs are saying is that
01:58 any regulation has to be done in line with European values of privacy and non-discrimination.
02:04 So in that spirit, they've effectively put off the table something called emotion recognition,
02:11 which is when computers can recognise or simulate human emotions. That now will fall under a
02:18 complete ban if, as I say, the trilogues then agree on the whole process. Similarly, predictive
02:25 policing comes under a near complete ban. You mentioned chat GPT in your introduction,
02:34 Rochelle. This is going to have its own bespoke set of rules. So those who actually make chat
02:40 GPT will have to disclose whether a text was originally generated by this method or not.
02:50 So disclosure and transparency is going to be very important. And on the question of
02:55 facial recognition, that's still actually a bit of a sore point because we saw today
02:59 in the parliament, the centre-right group, the European People's Party, tabled amendments
03:04 because they want to have some form of facial recognition as a useful tool for policing.
03:12 Other MEPs didn't want that. They think that's too much of an intrusion into people's privacy.
03:17 So those amendments didn't go through. But of course, there will be some member states,
03:22 particularly right-wing political forces, that will see artificial intelligence as a
03:27 useful tool for certain forms of policing. So they may try to push that particular agenda
03:32 going forward. As I said, and I stress again, this is not the end of the process. European
03:37 member states will have their say on this regulation of AI.
03:41 As you say, Armin, a long road ahead in this. But a first vote in any case. Armin, keeping
03:47 us up to date. Thanks so much.

Recommended