NGOs warn EU rules on AI need to do more to protect human rights

A group of 150 NGOs, including Human Rights Watch, Amnesty International, Transparency International and Algorithm Watch, signed a statement addressed to the European Union. In it, they urge the EU to not only maintain but also strengthen human rights protections when adopting AI legislation.

Between the algorithmic apocalypse and a cancer-free utopia, each side claims the technology could bring, but lies all kinds of pitfalls to avoid for responsible healthcare adoption. AI.

With Altman, Musk, and Zuckerberg among the first to jump into the black box, legislation is about to be enacted that will at least curb their enthusiasm. The European Union’s proposed law on artificial intelligence, or the AI ​​law, is the first legislation proposed by a major regulatory body. Two different camps claim to be either: Europetechnological sovereignty of, or b) not enough progress has been made in curbing the dangerous deployment of AI.

Transparency and Redemption

wednesday signer collective statement “Without strong regulation, companies and governments will continue to use AI systems, exacerbating mass surveillance, systemic discrimination, the concentration of big tech companies, irresponsible public decision-making, and environmental destruction. It will happen,” he warned.

This is not an “AI extinction crisis”. one liner statement. It contains certain parts of the Act that the authors believe should be maintained or strengthened. For example, the Accountability, Transparency, Accessibility and Remedy Framework includes publishing fundamental rights impact assessments, registering their use in publicly accessible databases, and ensuring that people affected by decisions made by AI. should include the obligations of AI adopters to ensure that they enjoy their rights. Right to be informed.

NGOs are also using AI-based public surveillance (e.g. Deployed during the coronation of King Charles). They call for a total ban on “real-time and post-mortem remote biometric authentication in publicly accessible spaces by all parties, without exception.” They also want the EU to ban the use of AI in law enforcement predictive and profiling systems, immigration status and emotion recognition systems.

Additionally, the letter’s authors tell lawmakers not to “succumb to the lobbying efforts of big tech companies to circumvent regulation for financial gain” and to decide which systems are classified as high risk. It seeks to uphold an objective process.

The bill divides AI systems into four tiers, depending on the level of risk they pose to health and safety or fundamental rights. Levels are unacceptable, high, restricted, and minimal.

High Risk vs. General Purpose AI

Applications such as social scoring systems used by governments are unacceptable, but systems used for things like spam filters and video games are considered minimal risk.

Under the proposed legislation, the EU would allow high-risk systems (such as those used in medical devices and self-driving cars), but impose restrictions on testing, documentation of data collection and accountability frameworks. Strict rules must be adhered to.

The original proposal did not include any reference to general-purpose or generative AI. However, in the wake of ChatGPT’s rapid rise last year, the EU approved a last-minute amendment that includes an additional section.

Business leaders have worked hard in recent months to influence the EU to water down the proposed text. They are particularly keen on what should be classified as high-risk AI, resulting in much higher costs. Some, like OpenAI’s Sam Altman, went on the offensive (throw) to appeal to personal charm a threat or two mixed in).

More than 160 executives, specifically from major companies around the world (including Meta, Renault and Heineken), also wrote to the committee. Initialization, they warned He said the bill “endangers Europe’s competitiveness and technological sovereignty”.

The European Parliament adopted a negotiating position on AI law on 14 June, beginning tripartite negotiations. These will require discussion among parliament, committees and the board before adopting the final document.

As this law was set with the aim of setting a global precedent (preferably a law that can evolve as technology evolves), Brussels is probably now, on behalf of all stakeholders, It will be crowded with enthusiastic defenders.