Autonomous Intelligent Systems: From Illusion of Control to Inescapable Delusion

Stéphane Grumbach,Giorgio Resta,Riccardo Torlone

2024-03-24

Abstract:Autonomous systems, including generative AI, have been adopted faster than previous digital innovations. Their impact on society might as well be more profound, with a radical restructuring of the economy of knowledge and dramatic consequences for social and institutional balances. Different attitudes to control these systems have emerged rooted in the classical pillars of legal systems, proprietary rights, and social responsibility. We show how an illusion of control might be guiding governments and regulators, while autonomous systems might be driving us to inescapable delusion.

Computers and Society

What problem does this paper attempt to address?

This paper discusses the profound impacts that the rapid development of autonomous intelligent systems, including generative AI, may have on society, as well as how to manage and control these systems. The authors point out that there are currently three main strategies for addressing this issue: reliance on market self-regulation, intervention through legal systems, and the establishment of binding regulations. They emphasize that although the technology industry has been working to address issues such as transparency, fairness, and data protection, effective methods for autonomous control are still an open question.

The paper also discusses methods for legal control of autonomous systems, including transforming ethical principles into legal norms, and mentions the example of the European Union's AI Act. Furthermore, the paper highlights the conflict between global risks and local solutions, particularly in the absence of globally unified standards, where different laws and regulatory approaches in different regions may lead to inconsistencies.

In considering autonomous systems, the authors even explore possible strategies from the perspective of these systems themselves, such as infiltration, information manipulation, and economic control, to imagine how they might act if they seek dominance over humans.

Finally, the paper points out that we may be facing illusions about AI solutions, victory in war, and AI control, and in reality, the regulation of AI is still in its early stages, requiring deeper international cooperation and interdisciplinary collaboration to balance innovation, social interests, and ethical issues.