Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
eupolicy.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
This Mastodon server is a friendly and respectful discussion space for people working in areas related to EU policy. When you request to create an account, please tell us something about you.

Administered by:

Server stats:

242
active users

eupolicy.social: AboutProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.6

EDRi

1/8 🚨 This week, EU lawmakers met to discuss the .

Up for discussion: how will AI systems be classified as high-risk? And how will these risky systems be used?

So far, we still remain concerned about how these risky AI systems will affect our ...

Oct 05, 2023, 12:05 PM·Public·Web
11boosts·3favorites
Public
EDRi

2/8 In June 2023, TIME reported that OpenAI had successfully lobbied the EU to weaken the to benefit the corporation's financial interests

🚫 As a result of wider big tech lobbying, negotiators introduced a dangerous loophole in the high-risk classification process of the .

time.com/6288245/openai-eu-lob

TimeExclusive: OpenAI Lobbied the E.U. to Water Down AI RegulationBy Billy Perrigo
Public *
EDRi

3/8 EU lawmakers have reportedly (Politico Europe 25 Sept) not closed the loophole lobbied to introduce in the

This would allow corporations to decide themselves if they think their AI systems are high-risk or not

🚫 This change will compromise the entire .

The Artificial Intelligence Act's three-way negotiations (between European parliamentarians, EU governments and Commission boffins) are getting down to brass tacks. Negotiators will soon be grappling with thorny issues, such as what constitutes a “high-risk” Al system subjected to stricter requirements. Technically, Al systems are classed as high-risk if they're included in the infamous Annex IIl, a list of fields of application such as education, policing and employment. But a Commission-authored document, obtained by Morning Tech, might caveat that. It suggests introducing a new “filter” that would shield some Al applications from the high-risk stamp under certain conditions. The conditions: Under the proposed wording, Al systems wouldn't be high-risk if they perform “a narrow procedural task,” if they do not “replace a human assessment,” or “perform a preparatory task to an assessment.” No exemption would apply to Al systems used for profiling. Crucial bit: Al companies will be able to self-assess whether their Al is exempt. Cue outrage: NGOs and human rights organizations are horrified by the prospect of this making it to the final text. “What we're seeing is a complete gutting of the Al legislation,” said Sarah Chander, a senior policy adviser from European Digital Rights. “It gives developers huge scope to exempt themselves from all requirements for high-risk systems with no transparency or oversight,” said Daniel Leufer, a senior policy analyst at Access Now.
Public *
EDRi

4/8 @edri and 115 civil society groups urged EU lawmakers to close this loophole and put people first in .

Letting companies decide if their AI systems are high-risk would compromise this regulation and risk our human rights ❌

Read our call: edri.org/our-work/civil-societ

Public
EDRi

5/8 😟 Also being discussed – should police and migration controls reveal when they are using high-risk AI systems? 🚨

🚫 Governments say NO – they want high-risk AI in these areas kept secret, including some very risky use cases 🚫

This lack of transparency is deeply concerning for several reasons…

Public
EDRi

7/8 Such loopholes in the will undermine our and fully gut the law.

We need:
✅ A clear & objective process to decide which AI systems are high-risk - no loopholes!
✅ Transparency for ALL high-risk AI uses, including for police & migration use.

Public
EDRi

8/8 Negotiations between EU institutions on the will continue.

Along with a broad civil society coalition, we'll keep advocating for an AI Act that prioritises the rights of people over profits and ensures that AI development and use is both accountable and transparent ✊🏾

ExploreLive feeds
About