On 20 October, the European Parliament became one of the world’s regulatory pioneers in field of AI when MEPs adopted proposals covering an ethics framework for AI, intellectual property rights and liability for damage caused by AI. According to the European Parliament, the recommendations would ensure that the EU becomes a ‘global leader’ in the development of AI.
Outlined below are the areas covered by the proposals recently adopted by the Parliament.
Spanish MEP, Iban García del Blanco’s legislative initiative, urging the Commission to propose a new legal framework defining the legal obligations as well as the ethical principles which AI developers should follow was adopted with 559 votes in favour, 44 against and 88 abstentions. Such framework would be adhered to all stages i.e. when developing, deploying and using AI, robotics and related technology in the EU (including algorithms, data and software).
MEPs decided that guiding principles should be integrated into future legislation such as respect for privacy and data protection, social and environmental responsibility, safeguards against bias/discrimination, right to redress, transparency and accountability as well as human-centric and human-made AI.
The initiative also contained provisions for high-risk AI (for example AI with self-learning capabilities) to be designed in a way that enables oversight by a human at any time. Furthermore, in circumstances where self-learning AI processes would result in a serious breach of ethical principles which could be deemed dangerous, a human should take full control and disable those processes.
Intellectual property rights (IPR)
A report authored by French MEP Stéphane Séjourné was adopted with 612 votes for, 66 against and 12 abstentions. The report stated that for the EU to be a global leader in AI development an effective IPR system is essential. In addition, it highlighted the need to protect developer’s innovation through adequate safeguards for the EU’s patent system. It also emphasised that in fostering such innovation neither the EU’s core ethical principles nor the interests of human creators should be sacrificed.
MEPs specified that AI should not be granted legal personality therefore only humans may own IPRs. They also drew a distinction between AI-assisted human creations versus solely AI-generated works. Also covered in the report were data collection, trade secrets, copyright, deep-fakes and use of algorithms.
Axel Voss MEP was behind the legislative initiative which establishes strict liability for those operating high-risk AI when use of such AI harms or damages life, health, physical integrity, property or that causes significant immaterial harm if it results in “verifiable economic loss”. This framework would provide businesses with legal certainty while also protecting individuals and promoting trust in AI.
Currently, high-risk AI technologies such as self-driving cars are rare, but MEPs backed proposals for operators of such AI technologies to have insurance just as operators of motor vehicles do. Mr Voss’ initiative was adopted with 626 votes for, 25 against and 40 abstentions.
A legislative proposal from the European Commission is expected early next year.