Proposal for a Regulation on Artificial Intelligence (AI)
Artificial Intelligence and the consequences of its rapid and limitless development are of concern to various sectors of society, including AI developers themselves. Without going any further, OpenIA’s own CEO, Sam Altman, recently appeared before the US Judiciary Committee to express not only the urgent need to regulate Artificial Intelligence, but in his own words stated ”My worst fear is that this technology will go wrong, and if it goes wrong, it can go very wrong”.
As if these words were not alarming enough, he continued his appearance on Capitol Hill by suggesting the creation of an international organisation to regulate and standardise the use and development of AI on an international level, equating the level of risk of AI to that of nuclear weapons.
What is the Proposal for a European Regulation on Artificial Intelligence?
The European Union is at the forefront of the commitment to regulate this technology, the European Commission published the White Paper on AI in 2020 where it already urged the urgency to address the challenges of complexity, unpredictability and autonomous behaviour of certain Artificial Intelligence systems.
Now, three years later, it has taken the next step by publishing the Proposal for a Regulation laying down harmonised rules on Artificial Intelligence (”Artificial Intelligence Act”).
This proposal for a Regulation will apply to providers marketing or operating AI systems on EU territory, irrespective of whether they are established in the EU or in a third country, to AI users located in the EU, as well as to providers and users located in a third country when the information generated is used on EU territory.
Furthermore, it introduces a definition of Artificial Intelligence that aims to stand the test of time and rapid technological change, defining it as a ‘software that is developed with one or more of the techniques and approaches listed in Annex I and can, for a given set of human-defined objectives, generate outputs such as content, predictions, recommendations, or decisions influencing the environments they interact with.’
What are the main issues to be introduced by the European Artificial Intelligence Regulation?
- Categorization of AI systems: An established mechanism is introduced to classify AI systems as: explicitly prohibited AI practices, high-risk AI, limited-risk AI and non-hazardous AI systems. Unacceptably dangerous AI systems, such as those designed to manipulate human behaviour, would be prohibited. High-risk systems, such as AI used in the health or transport sector, would be subject to strict requirements for transparency, oversight and accountability, as well as obtaining certification after notification to the relevant authority.
- Obligations for high-risk AI systems: High-risk AI systems would be subject to specific requirements. Suppliers of such systems would be required to conduct risk assessments, follow technical standards, ensure traceability of results and have contingency plans in place. In addition, there would be an obligation to carry out a pre-market conformity assessment and to keep detailed records of the training data used.
- Protection of Users’ fundamental rights: The draft of this Regulation emphasizes the need to protect and respect the principles of legality, transparency and non-discrimination. It expressly prohibits the use of AI for mass surveillance, social scoring, among others, as well as requiring the concurrence of qualified exceptions for the use of biometric recognition systems in public spaces. AI systems are required to be transparent and provide a clear explanation of how they work to end-users, avoiding opacity and allowing people to understand how AI-based decisions are made.
- Promoting innovation and ethical research: Despite the restrictions and obligations, the draft law also seeks to promote innovation and research in ethical AI. Measures will be put in place to facilitate access to high-quality data, collaboration between academic institutions will be encouraged, and funding programmes for ethical AI projects will be created.
How will the Artificial Intelligence Regulation be enforced and what sanctions can result from non-compliance or infringement?
In order to monitor and enforce the AI Regulation, cooperation between the European authority, the European Artificial Intelligence Board, as well as Member State authorities will be required. In this respect, in Spain, the national supervisory authority will be the Spanish Agency for the Supervision of Artificial Intelligence (AESIA), which will also have the sanctioning powers provided for in this Regulation.
Minor infringements, such as the submission of inaccurate, incomplete, or misleading information to notifying authorities, are subject to penalties of up to EUR 10 million, or 2% of annual worldwide turnover. For higher infringements a sanction of up to EUR 30 million or, if the offender is a company, up to 6% of the total annual worldwide turnover.
In short, this regulation will aim to establish a solid and ethical regulatory framework for the use of artificial intelligence (AI) in the European Union, and has been the subject of extensive debate and consultation, in order to achieve coherence with current data protection and consumer protection regulations, while at the same time carrying out the necessary balancing acts so as not to hinder the opportunities for innovation and economic growth that the development of AI entails.
While the Regulation, once adopted, will apply in the EU, it will inevitably have consequences outside the EU, and it remains to be seen whether it will influence the practices and regulations of third countries. It is worth noting that the G7 itself, which met in Japan in mid-May 2023, announced the creation of the “Hiroshima AI Process“, a collaborative forum to establish homogeneous international standards for the regulation of artificial intelligence.
At Letslaw we are experts in providing legal advice on digital law as well as Artificial Intelligence matters, so we will be glad to assist you in adapting your project to the new regulatory requirements on Artificial Intelligence. Do not hesitate to contact us!