EU's landmark AI regulation aims to both safeguard human rights and foster innovation
The European Union has reached a hard-fought agreement on a stricter, comprehensive regulation of artificial intelligence, news agency Bloomberg reported on Dec. 8.
"We have spent a great deal of time finding the right balance between maximizing the potential of A.I. to support law enforcement while protecting the fundamental rights of our citizens," Thierry Breton, the bloc's internal market chief, said, adding that members of the bloc do not want "mass surveillance" in Europe.
After more than 37 hours of negotiations held this week, delegates from the European Commission, the European Parliament, and the 27 member states agreed on a set of controls for generative artificial intelligence tools, such as ChatGPT from OpenAI Inc. and Bard from Google, capable of generating data after receiving the appropriate command.
The negotiators agreed to allow partial facial scanning in real-time, but with safeguards and exceptions, Breton said.
The agreement also bans biometric scanning that classifies people based on sensitive characteristics such as political or religious beliefs, sexual orientation, or race. Officials noted this as one of the most complex and delicate issues in the negotiations.
Bloomberg reports that the proposed law imposes financial sanctions on companies that violate the rules, with fines of up to EUR 35 million (approximately $37 million), or 7% of global turnover, depending on the violation and the size of the company.
The bill still needs official approval from the EU member states and the EU parliament. Despite this, the agreement marks a significant step towards a landmark A.I. policy that, in the absence of significant action from the U.S. Congress, will set the tone for the regulation of the rapidly developing technology.
The EU aims to introduce the first sustainable restrictions on A.I. outside of Asia. Politicians reportedly have been working for months to complete the formulation of the Artificial Intelligence Act and adopt it before the European elections in June 2024. This will lead to the creation of a new composition of the European Commission and the European Parliament, potentially introducing additional changes.
The decision was made at a session on Dec. 8, following a nearly 24-hour meeting that spanned from Dec. 6 to Dec. 7. During the initial meeting, some negotiators dozed off in the hall, while others discussed restrictions on real-time facial scanning in public before finally agreeing to suspend the negotiations.
Bloomberg notes that these lengthy discussions highlight how controversial the debate over artificial intelligence regulation has become, dividing world leaders and tech executives as the popularity of generative tools continues to grow.
The EU, like other governments, including the United States and the United Kingdom, has been trying to find a balance between the need to protect its own artificial intelligence startups, such as French Mistral A.I. and German Aleph Alpha, from potential societal risks. This proved to be a key stumbling block in the negotiations, as some countries, including France and Germany, opposed rules that they believed would harm local companies.
Some details will still be worked out by government officials in the coming weeks, but negotiators largely agreed to establish rules around generative A.I., which include basic transparency requirements for the developer of a large language model. It is noted that those who pose a systemic risk must sign a voluntary code of conduct to work with the commission to reduce risks.
Last spring, the parliament voted for a complete ban on biometric identification, but EU countries insisted on exceptions for national security and law enforcement. In the end, the parties agreed to restrict the use of the technology in public places using additional barriers.
Both the European Parliament and the European Council, which includes the 27 EU member states, must approve this agreement.
The United Kingdom, the United States, the EU, Australia, and China signed a joint declaration on Nov. 2, in which they recognized that artificial intelligence poses a potentially catastrophic risk to humanity.
Then, 28 governments of countries signed the Bletchley Declaration at the Artificial Intelligence Security Summit organized by the UK government.
The countries agreed to work together on artificial intelligence security research.
The UK’s Guardian newspaper noted that South Korea agreed to hold another such summit in half a year, and France in a year. However, it was noted that there is no international agreement on what a "set of rules" for artificial intelligence could look like or who should develop them.
The military war may be swinging in our favor, but the information war continues.
Just as an army needs soldiers, so does a free society need its journalists to ensure that people have access to honest, trustworthy voices to understand the world around them.
For the past five years, The New Voice of Ukraine has been working tirelessly to push back against Russian narratives and defend democracy. But we cannot do it alone.
Please consider supporting us on Patreon for just $5 a month – your donation does directly to supporting journalists and ensuring that this front of the infowar says solid and defended.
Thank you.
Follow us on Twitter, Facebook and Google News
