In January 2015, the JURI Committee of the European Parliament decided to establish a working group on legal questions related to the development of robotics and artificial intelligence (AI) in the EU, with a focus on civil-law aspects, writes Monica Monaco
In January 2017 this was followed by the adoption of a European Parliament Report with recommendations to the Commission on “Civil law rules on robotics”. The report calls on the Commission to propose EU legislation introducing a register of robots, setting up an EU Agency for Robotics and laying down principles of civil liability for damage caused by robots, which should be complemented by ethical codes of conduct.
Go deeper with GlobalData
The proposed code of ethical conduct would constitute the foundation for the identification, oversight and compliance with fundamental ethical principles of robotics, starting from the design and development phase.
In the European Parliament Report views, the code should not replace the need to tackle all major legal challenges in this field, but should complement regulation and should also introduce a procedure to allow robotics and AI to function in an ethically responsible manner.
While financial services use of AI is not directly cited in the Report, the Report opens a door to the potential development of any type of European regulation on AI, including in financial services. In the financial services sector algorithms are widely used and have wide effects.
Credit scoring algorithms raise or lower the scores of people based on external data; investment algorithms can cause big worries, as happened with the “Flash Crash” of May 2010, in which $1trn was wiped off the value of markets in no more than 10 minutes.
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Thank you!
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below form
By GlobalDataAlgorithms are AI, while it is difficult to determine from a legal perspective whether they are a service or a product.
Should AI be regulated as product or service?
Algorithms depend on external data sources and on pre-programmed routines and their behaviours correspond in the European Parliament view to some form of liability – contractual or extra contractual – which should entail some form of compensation.
Who would be liable for what and to whom?
According to the results of the 10 January 2017 European Commission consultation on data economy, which contained a chapter on liability related to the Internet of Things (IoT) technologies and autonomous systems, the most frequent damages suffered because of AI use are economic losses, both in terms of pure economic losses or losses linked to a missed opportunity.
The European Commission is consequently looking at whether to develop model contract clauses and best insurance practices in a specific sector for AI use, or in general every time AI is used. Furthermore, the Commission is looking at whether to develop standards for safety assessments and certification in a specific sector for AI use, or in general every time AI is used.
Should minimum liabilities be borne by certain service providers in certain sectors? A “risk-management approach” could, for example, allow that liability should lie with the entity most suited to manage the risks, which is often the producer.
Also, the possibility for human supervisors to interfere with AI is a factor which could play a role in the allocation of liability, as could event data recorders which save log-data.
Such changes could be implemented in European law via a general revision of liability law or via the expansion of the scope of Product Liability Directive (85/374/EEC) (PL) or even via the review of the Machinery Directive 2006/42/EC (MD).
The Machinery Directive is currently the object of a report, to be published in early 2018, which aims at checking whether the MD is fit for purpose to new emerging technologies as robotics or IoT.
Furthermore, the Commission has also conducted a public consultation on the evaluation of the PL Directive, seeking to assess the adequacy of existing liability rules when damages arise in the context of autonomous systems and advanced robots or IoT-systems.
However, Article 2 of the PL Directive explains that the text currently only covers movables and not software or digital services, so the Directive may need some “digitalisation”.
The European regulators motivation for regulating could arise from the protection of EU values, especially data protection, privacy and ethics, as well as from the will to avoid individual European member states developing regulation in isolation.
Is AI processing data for legitimate purposes under the General Data Protection Regulation (GDPR)? Are these data relevant and adequate for that very purpose?
Regulators could in theory have a say on the type of data that could be used and fed in AI, as well as on the role of the human supervisors’ interference with AI.
Based in Brussels for the past 14 years, Monica Monaco is the founder and managing director of TrustEuAffairs.
TrustEuAffairs advises its clients on Eu legislative relevant initiatives. Monica is a member of the Society of European Affairs Professionals (SEAP) since 2004, and served as a member of the SEAP Board from May 2012 to May 2015. Furthermore, Monica is Member of the Europol Virtual Currencies Taskforce and also Member of the European Commission Payment Systems Market Expert Group (PSMEG).
