News

Recent developments regarding the EU Framework for Artificial Intelligence (AI)

Artificial Intelligence is a wild beast with astronomical potential to benefit society, yet still a beast that must be carefully nurtured and tamed. With its innate characteristics such as its operational opacity and partially autonomous behaviour, AI poses potential risks for EU Citizens’ fundamental rights. Thus EU lawmakers are analysing the best ways to both embrace and regulate its use.

European Commission Proposal

The European Commission set out an AI strategy in 2018, with the goal being to reach EU-wide consensus on developing a ‘human-centric’ approach to AI that respects EU values and principles. In 2019, the Commission published a non-binding guideline on ethics in AI establishing 7 key criteria that AI developers should respect, including for instance; non-discrimination and fairness, accountability, transparency and privacy. In 2020, the Commission released a white paper that highlighted the importance of adopting a unified European approach and set the ball rolling for legislative proposals. Public consultation demonstrated that there was widespread agreement that the Product Liability Directive and national liability rules would need to be revised in order to cover risks created by the use of AI systems and to ensure compensation in the case of AI related damages. Moreover, there was general consensus that a new regulatory framework for AI is needed to complement the applicable legislation including those relating to consumer protection, data protection and privacy regimes. Importantly, the Commission is proposing to:

  • implement a prior conformity assessment for ‘high risk’ AI systems before entering the EU internal market, and
  • evaluate the IP framework to enhance access to and use of data.

European Parliament position

The Legal Affairs Committee (JURI) adopted 3 reports on AI on the 1st of October 2020.

  1. A framework of ethical aspects of AI, robotics and related technologies.

    The report recommends that the Commission establishes a widespread, future proof European legal framework of ethical principles for development, deployment and use of AI and related technologies in the Union. Moreover, the report recommends that the Commission should incorporate guiding principles focused on high-risk AI, robotics and related technologies, including non-discrimination and privacy considerations.  The Commission should further develop common guidance on the matter of high risk technologies, particularly by drafting a list of high-risk sectors that should help identify high risk AI technologies that are subject to compliance assessments. Ultimately, Parliament proposes that such technologies regarded as being compliant with the guiding principles are to be awarded with a European certificate of ethical compliance to be issued by national authorities.

  2. Civil liability regime for artificial intelligence.

    This report recommends the adoption of a harmonised legal framework for civil liability claims, including a new regulation listing high risk AI systems and the critical sectors in which they are used (to be updated every 6 months). The report proposes parallel systems of strict liability and fault based liability. On the one hand, operators of high-risk AI systems would be subject to strict liability for any damage caused by their AI operated systems or devices. Such operators must be subject to a mandatory insurance regime. Strict liability would also apply to AI systems, which, despite not being classified as high risk, repeatedly cause incidents resulting in serious damage. On the other hand, AI systems not listed as high risk remain subject to fault-based liability. The amount and threshold of compensation for damages, as well as procedural issues for bringing such claims should be determined by the EU.

  3. Intellectual property rights (IPRs) for the development of artificial intelligence technologies.

    The report recommends that an impact assessment is carried out by the Commission to assess the effects of AI and related technologies on the current legal regimes of patent law, trademark and design protection, copyright and ‘trade secrets’. EU laws should then be amended accordingly. The report holds that the Commission should encourage standardisation and create a balanced European data space to cater for the free flow, access, use and sharing of data while simultaneously protecting IPRs and trade secrets.

News update by Dr Gigi Gatt.

For more information please contact Dr Ian Gauci and Dr Terence Cassar.

Disclaimer: This article is not intended to impart legal advice and readers are asked to seek verification of statements made before acting on them.