The EU Parliament will vote on the AI Act on Wednesday, which aims to regulate the use of artificial intelligence. However, it will take another two to three years before this becomes noticeable in everyday life.
At least since ChatGPT, artificial intelligence has been on everyone’s lips. It is not new and we encounter it regularly in our everyday lives. At the latest when we unlock our cell phone with our fingerprint or face. Precisely because technology is being used more and more extensively, the EU has set itself the goal of creating a set of rules, the so-called AI Act. But only ChatGPT prevented this.
The EU Parliament will vote on the AI Act today, Wednesday. Now that so-called generative AI, which also includes ChatGPT, has been incorporated into the legislative proposal. Despite this delay, the EU is at the forefront.
With this law, which still needs to be coordinated with the EU Commission and the individual states of the community in the so-called trialogue, Europe would be a leader in the regulation of artificial intelligence (AI). Most states are still in the early stages of these efforts. Among other things, European regulations provide for a classification of AI software into risk classes, which bind vendors to certain requirements. Content created by programs such as ChatGPT must also be marked as such.
“We support an EU law that regulates the use of AI as soon as possible. The principle for this should be that we use the opportunities of this new technology and minimize its risks as much as possible”, say ÖVP MEPs Barbara Thaler and Lukas Mandl ahead of today’s vote on the so-called Artificial Intelligence Act in the plenary session of the European Parliament .
Green Thomas Waitz also describes the AI act as a “good model”. It clearly states what artificial intelligence can and cannot be used for: there are clear limits to social profiling, such as collecting and evaluating people’s biometric patterns or emotions.
It is “urgent and important” to provide a good framework for this rapidly developing technology. In the past, the EU has made the mistake of excluding risks but not exploring opportunities. The current proposal is balanced, even the industry is favorable, according to the representative of NEOS, Claudia Gamon. She is convinced that serious regulation can increase the potential of artificial intelligence.
keep reading
“Europe is at a crossroads when it comes to artificial intelligence. It will now be decided whether the AI Act will impede AI innovations in Europe and make it more difficult to use, or whether it will open up prospects for AI in Europe. Today’s vote in the EU Parliament does not end the discussion on the AI Act, it is entering its decisive phase. Now it must be a matter of correcting the weak points of the proposal in the final stretch of the legislative process”, warns Achim Berg, president of the industrial association of the German information and telecommunications industry.
How is the AI Act structured?
The law is comprehensive and applies to anyone offering an AI-based product or service. It covers applications that provide content, forecasts and recommendations or influence the user’s decision making. The focus is not just on commercial offerings, but also on the use of AI in the public sector, such as law enforcement. The “AI Act” intends to complement the General Data Protection Regulation (GDPR).
keep reading
Above all, AI programs used in critical infrastructure, law enforcement or education are considered “high risk”. Unlike “unacceptable” programs, they will not be banned. But you have to meet high requirements. This group includes, for example, software used for surveillance or that can manipulate photos and videos to be deceptively real (“deep fake”). Therefore, providers will likely have to record all of their AI’s activities and make this data available to authorities for review. This is associated with high costs.
A long way to the law
Assuming the EU Parliament agrees on a joint AI Act today, it will still be some time before this set of rules becomes noticeable in everyday life. Because, on the one hand, the trialogue negotiations only start when an agreement is reached, on the other hand, there are certainly generous transitional periods. For this reason, EU Technology Commissioner Margrethe Vestager has urged vendors to be required to carry out voluntary self-regulation. This could save time until legal rules come into effect, said EU Technology Commissioner Margrethe Vestager. After all, it will take two to three years for the AI Act to actually take effect. (catfish/APA/DPA)