The Intersection Between AI and IP: Conflict or Complementarity?

Abstract

Artificial intelligence (AI) is everywhere. If Alan Turing raised the question of providing machines with a form of intelligence as early as 1950, 1 AI has since revealed its potential thanks to big data and the improvement of computing power. The term ''artificial intelligence'' was popularized by John McCarthy and Marvin Lee Minsky, organizers of the 1956 Dartmouth conference that made AI a field of research in its own right. AI refers to systems that demonstrate intelligent behavior by analyzing their environment and taking action, with a degree of autonomy, to achieve specific goals. 2 ''Intelligence'' then refers to the fact that the machine imitates the cognitive functions associated with the human or animal brain, i.e. the ability to learn and solve problems. This may involve thinking or acting by imitating human behavior (cognitive approach) or rationality (computational approach). 3 If, for several decades, AI experienced waves of enthusiasm and setbacks (''winters'' of AI), it is now part of everyday life, whether it is using a personal virtual or virtual assistant, or travelling in a semi-autonomous vehicle. It is also generating a craze for solving or anticipating complex problems in medicine, policing or justice among others. The adaptability of AIs greatly increases their applicable uses, thus different objectives can be pursued: automatic natural language processing, knowledge representation, automated reasoning, machine learning, computer vision and robotics to name a few. In this last case, AI systems are not only working on the software level, but they are also acting in a virtual world integrated into hardware.

Mitigating the adverse effects of AI with the European Union's artificial intelligence act: Hype or hope?

“…It appears that the AIA prioritizes economic, business, and innovation over moral concerns and that human rights are just an afterthought (Castets‐Renard & Besse, 2022). Noteworthy is that fundamental rights seldom appear in the main text; the primary focus of the AIA is clearly on markets and market access of AI systems.…”

“…As such, the proposed approach leaves the preliminary risk assessment, including identifying AI systems as high‐risk to the provider and developer. The latter possess consequently a significant amount of discretionary leeway (Smuha et al., 2021), for example, they can decide whether the used software is an AI system, whether the system may cause harm, and how to comply with the mandatory requirements of Title III, Chapter 2 AIA (Castets‐Renard & Besse, 2022); furthermore, they can (theoretically) classify high‐risk technologies as adhering to the rules using the self‐assessment procedure (EPRS, 2022a). Critics fear the rules on prohibited and high‐risk AI practices may prove ineffective because the risk assessment is left to the provider/developer (Kop, 2021).…”

“…The delegation or outsourcing of rule‐making power to European standardization organizations (i.e., standardization and notified bodies) might also lead to a lack of democratic oversight and the impossibility for relevant stakeholders, such as civil society organizations and consumer advocacy groups, to influence the drafting of standards as well as the judicial means to monitor and control them once they have been adopted (Bansal et al., 2023; Castets‐Renard & Besse, 2022; Ebers et al., 2021). Lack of stakeholder participation : Standardization processes and the related bodies often lack adequate involvement of all relevant stakeholder groups, including consumer and civil society representatives (Castets‐Renard & Besse, 2022). Especially consumer organizations often struggle to participate in such processes, for example, due to a lack of financial resources, expertise, experience, or political representation.…”

“… Power asymmetries : Standards are usually developed by standard‐setting organizations in which industry representatives exert significant influence, whereas the voices of civil society and consumer advocates are often overheard (note that, contrary to business organizations, civil society organizations often lack E.U.‐level representation). In the context of AI, there are also significant power imbalances between AI providers and developers, on the one hand, and other relevant stakeholder groups and standardization and notified bodies on the other (Castets‐Renard & Besse, 2022). Lack of transparency : Little is known about the activities of standardization and notified bodies, mainly due to the frequent outsourcing of testing, inspection, and certification activities to third parties; critics accuse those organizations of relying too much on corporate self‐assessments instead of conducting independent and impartial monitoring (Veale & Zuiderveen Borgesius, 2021).…”

“…Only those with obligations under the AIA can challenge regulators’ decisions, not those whose fundamental (human) rights are negatively impacted. Given the lack of adequate complaint and judicial redress mechanisms, the AIA lacks a “bottom‐up force to hold regulators accountable for weak AIA enforcement,” or to put it differently: the AIA lacks effective means of private enforcement (i.e., individual enforcement rights) to compensate for the ineffective public enforcement (Castets‐Renard & Besse, 2022; Ebers et al., 2021; EPRS, 2022a; Smuha et al., 2021; Veale & Zuiderveen Borgesius, 2021).…”