European AI Law

The researcher at the Valencian Research Institute for Artificial Intelligence (VRAIN) at the Universitat Politècnica de València (UPV) and member of the Joint Research Unit of the Valencian School of Postgraduate Studies and Research Network in Artificial Intelligence (ValgrAI), José Hernández-Orallo, has been selected to join the Scientific Committee that will advise the Office of Artificial Intelligence, established by the European Commission and comprising 60 international experts.

This Scientific Committee will advise on the implementation of the European AI Law and the assessment of the impacts and risks of general-purpose AI models (IAPG).

A leader in AI evaluation

José Hernández-Orallo is a researcher at VRAIN at the UPV and a Full Professor at the Universitat Politècnica de Valencia; he is the driving force behind the AI Evaluation Programme at the ValgrAI Foundation and is currently also Director of Research at the Leverhulme Centre for the Future of Intelligence at the University of Cambridge (United Kingdom).

His academic and research work spans different fields of artificial intelligence, machine learning, data science and the measurement of intelligence. It focuses on a more in-depth analysis of the capabilities, generality, progress, impact and risks of artificial intelligence.

Since 2017, he has led the scientific agenda for AI evaluation through new measurement paradigms and evaluation frameworks published in Nature, Science and at leading international AI conferences. He is the founder of aievaluation.substack.com and ai-evaluation.org.

The main responsibilities of the AI Law Scientific Committee include alerting the AI Office to systemic risks, advising on the IAPG's classification and evaluation methodologies, and supporting activities to monitor the deployment of AI in the EU.

This Scientific Committee complements the Advisory Forum, which provides the Commission and the AI Council with broader input from stakeholders in industry, civil society and academia.

Geographical balance and gender parity have been considered in selecting expert members. José Hernández-Orallo has been selected alongside two other experts from Spain: Javier Del Ser and Asunción Gómez-Pérez. They will be appointed to the panel for a renewable term of 24 months, in a personal capacity.

General-purpose AI

Specifically, José Hernández-Orallo, researcher at VRAIN at the UPV, explains that "the panel will focus primarily on general-purpose AI with systemic risks – that is, AI which, due to its high capabilities, scope or novelty, may pose a systemic risk, meaning a threat that could affect societies or economies on a global or large scale".

In this regard, the panel will provide advice through short-, medium- and long-term working groups, whilst also carrying out constant monitoring, with the ability to issue alerts to the European AI Office regarding the risks posed by new models or phenomena of concern due to their widespread use.

José Hernández-Orallo also states that, for his part, "I will endeavour to bring scientific rigour to the assessment of AI capabilities and tendencies, shifting the panel's work more towards anticipating the risks posed by frontier systems and future superintelligence, to complement the work on information and mitigation".