In addition to highlighting the need to establish proper risk-assessments and risk-based approaches (evoking the “high risk” perspective of the European Commission White Paper), the text highlights the need for proper safety features, to ensure both the transparency and accountability of AI systems. Such systems should also not be biased or discriminatory, so as to guarantee full respect of fundamental rights. In that sense, the EP considers that AI should not increase discrimination and bias, rather that it should aim to ensure social responsibility and gender balance, issues currently at the core of numerous European societies. In the same vein, AI should also help in terms of the quest to respect the environment and sustainability. 

The text also provides interesting and important guidance on the respect of privacy, especially where biometric recognition systems are concerned. In this respect, EP underlines that good governance should be ensured, particularly in situations where consumers might be the target of algorithmic systems.

Moreover, the EP emphasizes that some additional areas, such as security and defence, transport, employment, workers’ rights, digital skills, education, culture and the national supervisory authorities, should be the object of particular and sectoral focus. 

Finally, the text highlights the need for both European coordination and international cooperation, and underlines the possibility of establishing European certification of ethical compliance in order to enable better harmonization among EU Member States.

• Civil liability Regime for Artificial Intelligence

The resolution on a civil liability regime for AI was mainly based on Axel Voss’s report and followed the release of the European Parliament Research Service study on civil liability for AI at the end of September 2020. In this resolution, the EP affirms that there is no need for a complete revision of the ‘‘well-functioning liability regimes’’ currently in place among EU Member States. Nevertheless, in highlighting the ‘‘complexity, connectivity, opacity and vulnerability, the capacity of being modified through updates, the capacity for self-learning and the potential autonomy of AI-systems, as well as the multitude of actors involved’’, the EP reiterates that civil liability is one of its main challenges.

The resolution acknowledges that the ‘‘existing fault-based tort law of Member States offers in most cases a sufficient level of protection for persons’’. However, due to the complex, intractable and impecunious nature of the third-party, additional guidelines are still needed. Thus, the resolution focuses on civil liability claims against the operator of the AI-system: it affirms that the operator’s liability is justified by the fact that they are responsible for the risk associated with the AI system as the operator might be, in many cases, the first point of contact for the victim. 

The resolution also highlights the fact that various liability rules should be applied and be based on a risk-based approach. However, clear criteria and definitions are still currently being discussed with the intention of ensuring a high level of legal certainty within the EU.

The final part of the text reiterates one of the main considerations raised by European Commission’s White Paper, as it highlights the crucial role of insurance companies and the need for appropriate coverage of algorithmic and AI-based systems.

• Intellectual property rights for the development of artificial intelligence technologies

The last resolution, based on Stéphane Séjourné’s Report focuses on the intellectual property rights at stake in the development of AI technologies, while stressing that this topic has not been tacked by the European Commission yet.