How to deal with the hype, ethical complexity and sustainability challenges
First published in AI Business, September 2021.
The European Commission published in May a proposal for a regulation on artificial intelligence (AIR), the first-ever legal framework to govern the use and development of AI. AIR “supports the Union’s objective of being a world leader in the development of safe, reliable and ethical artificial intelligence […] and ensures the protection of ethical principles.
The draft regulation establishes “an ecosystem of trust” in which the development of AI is promoted while respecting citizens’ rights under the EU Charter of Fundamental Rights. The Commission wanted to anticipate and prevent the risks associated with the use of this new technology. This is a direct attempt to raise the bar on existing AI standards, as the EU has already done in other sectors, especially in the life sciences.
No industry is more familiar with the technological hype, ethical complexity and sustainability challenges than the field of life science regulation. Just think of IVF, cloning, human embryonic stem cells, hybrid and synthetic embryos, GMOs, editing of the human genome: each of these innovations has encountered ethical and social challenges, and therefore , policies and regulations.
Under the proposed regulation, companies behind artificial intelligence products will be required, for the first time, to deal with notified bodies and apply for authorizations; they can even see the scope of their limited product usage. This will sound familiar to those with experience in the medical devices and diagnostics industry, who will be well used to the challenges of demanding regulation and having to prove compliance.
For others, it will be a shock.
Lessons to learn
No field is as complex in data as the life sciences, so it’s no surprise that AI is making such dramatic advances there. Medicine in particular has a lot to gain (think protein), but with lives involved, this transformative power, with its reliance on machine-driven assumptions, needs to be triggered with caution.
What lessons can we learn from previous attempts to regulate technological change?
Life sciences lawyers are almost spoiled for choice when it comes to examples of past ambitious regulatory “raising the bar” efforts that have not gone quite as planned. The EU’s Medical Device Regulation (MDR), for example, became fully applicable on May 26, 2021, after four years of market turmoil due to new requirements to obtain a CE mark.
Like AIR, the MDR required re-certification of almost all products. The MDR also required the requalification of all notified bodies, the organizations designated to assess the conformity of products before they could be placed on the market, which predictably created bottlenecks and delays.
Like the MDR, the In Vitro Diagnostic Medical Device (IVDR) Regulation, which is expected to enter into force in the EEA in May 2022, requires most in vitro diagnostics to obtain a CE mark with the participation of a notified body, whereas previously manufacturers were able to self-certify. Notified Bodies, who are just recovering from all MDR applications, are already facing a heavy new workload, with industry associations1calling for a delay in the implementation of the IVDR.
AI regulators should learn from the MDR / IVDR experience: What looks good on paper can turn into a practical nightmare if the bureaucracy is too heavy and the infrastructure too weak. Will there be enough expertise to fill the job description of an AI Notified Body? Can we avoid another regulatory bottleneck?
Another lesson to be learned from the life sciences industry is to avoid putting regulations in place too early, when they can become counterproductive, like the EU’s GMO directives.
These were designed to protect human health and the environment from the theoretical dangers of organisms produced using recombinant DNA techniques, but ended up super-bureaucratizing the use of precisely edited organisms (per example with CRISPR Cas-9 technology, winner of the Nobel Prize), while exempting the production of random mutants by ionizing radiation (for example in the livestock sector).
Failing to protect the environment or human health, the GMO directives were simply too cautious, hampering European research and competitiveness at a time when gene editing technologies have a vital role to play in addressing global challenges. food security, climate change, biodiversity, health, energy consumption and sustainability.
The right balance
Fortunately, the European Commission is well aware of the problem of gene editing, and has recently reassessed the scheme, concluding “that there are strong indications that the applicable legislation is not suited to certain objectives for some. [new genetic technologies] and their products, and that it must be adapted to scientific and technological progress. Given the potential of artificial intelligence, the hope is that they see the parallels.
The Commission has boldly asserted that AIR will reconcile rights protection and make the EU a world leader in the development of AI. This is a crucial balance to achieve.
If all the weight is on the protection of rights, it can fall on the shoulders of innovative companies, in the form of the regulatory burdens of development. In other words, the cost of excessive regulation can deter small innovators, leading to a market dominated by giant companies.
It is not impossible to find the right balance and the Commission believes so.
In fact, the RIA is almost pedantic based on the ‘New Legislative Framework’ (NLF) regulatory model, under which notified bodies undertake product conformity assessments, a proven approach for many product areas. , including medical devices and IVDs. . Additionally, to allow smaller innovators to enter the industry, the RIA includes provisions for “regulatory sandboxes” to foster the development of AI systems in a controlled environment for a limited time.
It seems that the European Commission has learned a lesson from the experience of GMO directives: protect against risks, but encourage innovations that can benefit the environment, human health and the economy. Hopefully they learn from the life science industry and translate the project into a lean and efficient system.
1. In particular the European Federation of Pharmaceutical Industries and Associations, and the European Coalition of Cancer Patients
The content of this article is intended to provide a general guide on the subject. Specialist advice should be sought regarding your particular situation.