Brussels, 15 June 2022
On Wednesday MEPs of the EP’s Culture and Education committee adopted their recommendations for the upcoming AI Act, to be included in EP’s position for legislative negotiations the Council.
The Culture and Education committee asks for increased transparency obligations for those AI technologies, which are used to create or disseminate machine-generated news and to recommend cultural and creative content. The users need to be informed “in a timely, clear and visible manner” about parameters used for the moderation of content and personalized suggestions.
Proctoring systems to be treated as high-risk
They demand to extend the scope of educational AI technologies that shall be treated as high-risk. Systems used to monitor and detect student’s behaviour during tests to detect cheating (so-called e-proctoring systems) must be included, as well as technologies used for determining the areas of study a student should follow.
Quote
The rapporteur Marcel Kolaja (Greens/EFA, CZ) welcomed the agreement that the e-proctoring systems should be among the high-risk applications.
“It makes a huge difference whether you pass the exam or not. Therefore, we cannot leave it up to technology without proper rules and assessments to decide and determine our lives. While its name may suggest otherwise, artificial intelligence is far away from perfect. It can discriminate students based on their gender or skin color. If your skin is a bit darker than the type of skin that the dataset was trained on, it may falsely report that you are holding a dark object in your hands. The system then may automatically report you as a suspect of cheating. Those situations must be avoided“, he said.
AI-literacy
CULT MEPs also ask to include special provisions to oblige the EU and the Member States to promote development of a sufficient level of AI literacy, through education and training, skilling and reskilling programmes. Such literacy should include basic notions and skills about AI systems and their functioning, different types of products and uses, their risks and benefits and the severity of the possible harm they can cause, they say.
Background
The AI Act, which is considered a main pillar of the EU digital single market strategy, will set out horizontal rules for the development, modification and use of AI-driven products, services and systems within the territory of the EU, applicable to all industries. It aims to establish an EU-trustworthy AI paradigm, which requires AI to be legally, ethically and technically robust, while respecting democratic values and fundamental rights.
The Commission’s proposal was presented 21 April this year. The EP committees on Internal Market and the Civil Liberties which are the lead committees on the legislation will vote on their recommendations in late October, and in plenary in November.