Along with the green transition, digitisation is the highest strategic priority of this mandate at the European institutions given its link to the sustainability of the European welfare and development model. The European Parliament, Commission and Council are currently developing an ambitious regulatory agenda for the digital transition. This agenda, on the one hand, involves the updating of the now obsolete rules in e-commerce with the Digital Services Act. On the other, it encompasses the adoption of a completely new regulatory package to civilise the digital environment -while ensuring and promoting our sustainable development model and rights in a global environment (Digital Market Act, Data Act, Data Governance Act…). The European Parliament is presently working on the Artificial Intelligence (AI) Act that the European Commission unveiled in April 2021. As many as five parliamentary committees have an opinion or exclusive competence on some parts of the regulation, and this work should result in a European Parliament proposal that will be negotiated with the Council later this year. The European Union will thus be the first in the world to have a comprehensive/holistic regulation encompassing the AI phenomenon.
There is no time to waste. The massive deployment of AI is a technological leap that will bring about (and is already causing) substantial changes in the labour market, in the relationship with public authorities, in personal relationships, and even in our domestic life. At the same time, its development will entail major differences in global competitiveness. For the EU, the challenge remains in finding the right balance between protecting citizens’ fundamental rights, and boosting investment, research and the implementation of AI systems. We must promote technological development without hindrance while creating an ecosystem of trust for European citizens. The ultimate goal of AI can only be the improvement of our societies and the lives of our citizens.
Any technological progress entails benefits and risks. When Dworkin spoke of science, he highlighted its ambivalence both as a promise and as a threat. With regard to AI, these benefits and/or risks are on a scale hitherto unknown, given its intrinsic potential. AI will will take us through dizzying changes that would otherwise take generations. Fortunately, AI provides us with powerful tools to better approach the major challenges of our time: the fight against climate change and depopulation -even anticipate future pandemics such as COVID19, which has put our resilience as societies on the ropes, and that we can now better address by accelerating research on drugs, vaccines, and developing state-of-the-art monitoring applications. Nevertheless, because of this potential, a controlless AI can also cause risks and damages on a scale and with a speed previously unknown. Regulating AI in Europe is a matter of ethics.
I had the honour myself of being the rapporteur for the first European Parliament legislative initiative on the “Ethical Aspects of Artificial Intelligence, Robotics and Related Technologies” adopted in October 2020 with an extraordinary support of the majority of the plenary of the Parliament. It was meant to indicate to the Commission our ideas and positions regarding the future Regulation. The 2021 AI Act claims to be inspired by the Parliament’s proposal, but it falls far short in addressing the ethical dimension. In its text, the Commission has largely replaced citizens by market and this is something that we must correct in the Parliament’s position.
The European Union wants and can be a pioneer in the legal establishment of an ethical threshold that provides an added value of trust to the European AI in the world, and protects European citizens from the possible adversities that this technological evolution entails. This ethical threshold must be consistent with our European principles and values, as reflected in the Charter of Fundamental Rights of the European Union, and must be congruent with our civilising project. It must be a legislation inspired by a humanist approach and focused on a technological development, applicable not only to AI developed in Europe, but also to AI systems that want to operate in the Union. It should aim to become a shared basic global framework on the minimum requirements for the development and use of this technology. We pursue that the so-called “Brussels effect”, which worked so well with the GDPR, will also serve to globalise our demanding civilisational standards.
However, the development of AI cannot stay put in the offices of the institutions or expert circles. We must democratise knowledge of AI in order to ensure that the human being is at the centre of the equation. Sector planning should not be left solely in the hands of the market. We assert that the need for public participation that safeguards objectives beyond possible economic profitability, so that the measurable aspects in terms of social profitability are also the subject of research and development: better public services, social responsibility, environmental sustainability, gender equality… The objective is also that the implementation of this regulatory framework engages all citizens, especially the individuals and groups most involved or affected.
In this vein, the text approved by the European Parliament in 2020 established a specific mandate for all European and national supervisory bodies to regularly and mandatorily involve civil society in the drafting of the regulatory governance model −paying special attention to the perspective of small and medium-sized enterprises. We were seeking both, a co-responsibility in the implementation and execution, and the inclusion of civil society, social partners and consumers, in one way or another, in the design of the governance mechanisms. My aim is to recover provisions in this sense in the final text.
AI is a gaseous matter, a reality in permanent and rapid evolution, even capable of improving itself (or worsening itself, depending on the values at stake) with a certain degree of autonomy. We must be aware of not to establish a rigid regulatory framework that becomes obsolete with the arrival of next innovation. Besides, one of the main objectives of regulation is to add critical mass, avoiding fragmentation of the European digital internal market −while complying with the sacred principle of subsidiarity. Thus, the 2020 European Parliament’s resolution designed a top-down governance model: a European coordinating body, able to harmonise legislative development across the Union, and to adapt quickly to technological progress. It also resulted in entities in charge of administering and enforcing regulations at the national level. The AI Act falls short, as it proposes a coordinating body for Member States at the European level (mirroring the GDPR model of governance; one of the aspects with most room for improvement), and delegates everything else to national authorities. In my opinion, a European Agency must ensure the unity of the model −one with sufficient tools to adapt to changes with agility, avoiding the need to over-regulate every bit that generates this technological environment in constant metamorphosis. Leveraging on the debate on the new Data Act and the symbiosis between AI and data, both rules should be the genesis of a powerful European Agency to deliver on the aforementioned objectives. This agency should also support operators, in particular SMEs and start-ups in complying with the rules, help to the global harmonisation of basic regulatory standards and disseminate knowledge of AI literacy to the public..
As noted above, the general public needs to enter into the debate and be included in the understanding of the profound implications of AI. Such a disruptive technology has to be subject to democratic control; societies at large have to be aware of this reality and be able to make informed decisions about the model of society they want. It is therefore preferable to speak of “digital literacy” rather than the “acquisition of digital skills”, or even “digital alphabetisation”, when talking about the adaptation of society to this new digital environment. I believe that it is the duty of public authorities to make the main features of this process accessible to citizens, as well as it is our duty to ensure a mature public debate on AI. The very survival of democracies is at stake here, given that we have just begun to witness the disturbing social effects produced by the uncontrolled application of AI.