Breakthrough for regulating AI and banning discriminatory technologies
Artificial Intelligence Act
Today, Members of the European Parliament have just voted on the world’s first law regulating Artificial Intelligence. This will determine the Parliament’s position for the trilogue, which is due to start this evening. Thanks to the Greens/EFA Group, the AI Act will include safeguards on environmental standards and fundamental rights linked to the deployment of AI systems on the market. MEPs voted to fully ban biometric mass surveillance in Europe’s public spaces.
Kim van Sparrentak MEP, Greens/EFA shadow rapporteur in the IMCO committee, comments:
“Artificial Intelligence has enormous potential for both incredible developments and dangerous risks to humanity. We need to ensure that AI works for people, society and the planet, and not for surveillance states or tech capitalism.
“We welcome that the EU is leading the world in the regulation of AI to ensure that it benefits our society. The European Parliament is the first to call for strict and binding rules for models like ChatGPT or Dall-E. Europe needs to take back the reins to ensure these systems with an enormous impact on society also need to comply with the rules, so they are trustworthy, safe and unbiased.
“With AI and the data centres they run on guzzling energy and increasing our emissions, we must ensure AI promotes, and does not hamper the green deal. Thanks to the efforts of our Group, we have included key environmental provisions that will enable better monitoring of the environmental impact of systems, design requirements and reporting obligations.”
Sergey Lagodinsky MEP, Greens/EFA shadow rapporteur in the LIBE committee, comments:
“Our Group has been consistently advocating for a human-centric approach to AI. We want to put fundamental rights at the core of AI legislation and we have succeeded in adding multiple obligations, including a fundamental rights impact assessment, to the text, which will have a significant impact on how systems are developed and used in the EU.
“A number of bans will prohibit the most unethical and undemocratic practices in the EU, including biometric mass surveillance, AI systems used to recognize emotions, gender or sexual orientation or AI used for predictive policing. These applications of AI are in use already in countries like Russia and China with a detrimental effect on fundamental rights and civil liberties. All of these bans were not part of the Commission proposal and are a big win for our group. By and large, the compromise found achieves a reasonable balance between responsible regulation and appropriate incentives for innovation."
The AI Act is on the way to be the world’s first legislation on artificial intelligence and contains harmonised rules on the placing on the market of such systems. It distinguishes between 3 categories of AI systems, prohibited, high risk and low-risk systems, and contains specific obligations for foundational models (such as “ChatGPT”).
The Greens/EFA Group has long been campaigning for a ban of facial recognition technologies, which wrongfully report large numbers of innocent citizens and systematically discriminate against under-represented groups.