The EU AI Act, which is set to come into force on the 1st August 2024, marks the European Union’s (EU) first piece of regulation related to artificial intelligence (AI). With the first regulatory framework for AI proposed by the European Commission in April 2021, providers and users are now facing down lawful requirements they must abide by.
Act looks to protect rights
The Act, which was officially adopted by the European Parliament in March 2024, seeks to protect rights, democracy, law and environmental sustainability from AI it regards as high-risk. It sets out obligations for AI depending on its risks and level of impact.
“We finally have the world’s first binding law on artificial intelligence, to reduce risks, create opportunities, combat discrimination, and bring transparency,” said Brando Benifei, Internal Market Committee co-rapporteur. “Thanks to Parliament, unacceptable AI practices will be banned in Europe and the rights of workers and citizens will be protected. The AI Office will now be set up to support companies to start complying with the rules before they enter into force.”
“The EU has delivered. We have linked the concept of artificial intelligence to the fundamental values that form the basis of our societies,” added Dragos Tudorache, Civil Liberties Committee co-rapporteur.
Banned and monitored AI systems
Unacceptable AI systems considered a threat, will include cognitive behavioural manipulation of people or specific groups; social scoring whereby people are classified; biometric identification and categorisation of people; and real-time and remote biometric identification systems such as facial recognition.
Exceptions may be allowed such as real-time and remote biometric identification systems being allowed in instances where the case is regarded as serious.
High risk AI systems, meanwhile, will be classified as negatively affecting safety or fundamental rights, and will be divided into two groups: AI systems used in products and AI systems that need to be registered into an EU database such as law enforcement and education. In this scenario, high-risk AI systems will be assessed prior to market entrance and throughout their lifecycle.
Crucially, Generative AI, which includes ChatGPT, will not be considered high-risk but will need to comply with transparency requirements and EU copyright law. It must disclose the content was generated by AI; design the model to stop it from generating illegal content; and publish summaries of copyrighted data used for training.
The latter point has proved particularly divisive, as artists, authors and actors have been vocal in recent years as what they regard as the threat AI poses. Fears circulate around AI models being trained on their voices and faces to create content without being properly compensated for it.
In recent news, Hollywood performers have gone on strike over the use of AI, as the Screen Actors Guild American Federation and Radio Artists (Sag-Aftra), which brought Hollywood to a standstill in striking last year, has called for industrial action. Their concern is that gaming studios will use generative AI to create video game characters.
Businesses that fail to comply with the EU AI Act will face punitive measures. They could be fined up to €35,000,000 or 7% of their annual turnover.
“The AI Act will introduce limited exceptions for text and data mining and recognise the importance of balancing copyright protection with promoting research and innovation. It acknowledges the need for proportionality in compliance requirements for startups and SMEs,” said Sebastian Gierlinger, VP of Engineering at Storyblok. “The AI Act requires transparency from providers, ensuring accountability and enforcement of copyrights. AI companies will be required to provide comprehensive information about the datasets used.
“With the implementation of the AI Act, companies must develop a comprehensive AI policy that serves as a framework for responsible and transparent AI deployment. It is important to have an AI policy that ensures that the technology is used ethically, legally, and effectively.”
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by commenting below or visiting our LinkedIn page.
