EU guidelines on AI should do extra to guard human rights, NGOs warn

A group of 150 NGOs including Human Rights Watch, Amnesty International, Transparency International and Algorithm Watch have signed a statement addressed to the European Union. In it, they call on the Union not only to maintain the protection of human rights when passing the AI ​​law, but to improve it.

Between the algorithmic apocalypse and the cancer-free utopia that various camps claim the technology could bring lies a spectrum of pitfalls that must be avoided for the responsible use of AI.

As Altman, Musk, Zuckerberg and others fall headfirst into the black box, legislation aimed at at least curbing their enthusiasm is on the way. The European Union’s proposed artificial intelligence bill – the AI ​​Act – is the first of its kind by a major regulator. Two different camps claim that this either a) paralyzes Europe’s technological sovereignty or b) does not go far enough to curb the dangerous use of AI.

transparency and redress

Signatories to Wednesday’s joint statement warn: “Without strict regulation, companies and governments will continue to deploy AI systems that exacerbate mass surveillance, structural discrimination, centralized power of big tech companies, irresponsible public decision-making and environmental damage.”

The <3 of EU technology

The latest rumors from the EU tech scene, a story of our wise founder Boris and questionable AI art. It’s in your inbox for free every week. Join Now!

This isn’t “AI faces extinction” one line statement. It contains certain parts of the law that the authors feel need to be retained or improved. For example, a “framework for accountability, transparency, accessibility and redress” must include the obligation for AI users to publish impact assessments on fundamental rights, register use in a publicly accessible database and ensure that people affected by AI decisions , this can also do Right to Information.

The NGOs also take a clear stance against AI-based public surveillance (like the used during the coronation of King Charles). They call for a total ban on “real-time and post-remote biometric identification in public spaces by all actors without exception.” They also call for the EU to ban AI in predictive and profiling systems in law enforcement, as well as in migration contexts and emotional recognition systems.

In addition, the authors of the letter call on lawmakers to “resist the lobbying efforts of large tech companies seeking to circumvent regulation for financial reasons” and to maintain an objective process for determining which systems are deemed to be at risk.

The proposed law will classify AI systems into four tiers based on the level of risk they pose to health and safety or fundamental rights. The levels are: unacceptable, high, limited and minimal.

High risk vs general purpose AI

Applications like social scoring systems used by governments are unacceptable, whereas systems used for things like spam filters or video games would be considered minimal risk.

Under the proposed legislation, the EU will allow high-risk systems (e.g. those used for medical devices or autonomous vehicles), but operators will have to adhere to strict rules around testing, data collection documentation and accountability frameworks.

The original proposal made no reference to general or generative AI. However, after ChatGPT’s meteoric rise over the past year, the EU approved last-minute changes to include an additional section.

Business leaders have been working hard in recent months to persuade the EU to water down the proposed text. They placed particular emphasis on AI, which should be classified as high-risk AI, resulting in significantly higher costs. Some, like OpenAI’s Sam Altman, launched a personal charm (throw) offensive. a threat or two in the mix).

Others, notably more than 160 executives from major companies around the world (including Meta, Renault and Heineken), have also sent a letter to the commission. In it, they warned that the draft law would “endanger Europe’s competitiveness and technological sovereignty”.

The European Parliament adopted its negotiating position on the AI ​​law on June 14 and the trilogue negotiations have now begun. This requires discussions between the Parliament, the Commission and the Council before adopting the final text.

As the law aims to set a global precedent (though hopefully one that can evolve with technology), Brussels is likely teeming with concerned advocates right now – on behalf of all interested parties.

Comments are closed.