果冻传媒

In his PhD. research, raises the question of whether artificial intelligence (AI) can become more ethical than humans. This inquiry stems from literature on artificial superintelligence, which explores the possibilities and plausibility of AI surpassing human capabilities. The answer to this question, according to Cannon鈥檚 research, depends on fundamental assumptions about the nature of human and machine cognition.

Two models

Cannon identifies two models in thinking about the mind: 鈥渃ognitivism鈥� and 鈥減ost-cognitivism.鈥� Cognitivism is the dominant model, upon which existing AI research heavily relies. In this model, the mind is viewed as a kind of machine, akin to a computer. Cognition and intelligence are seen as problem-solving abilities.

On the other hand, post-cognitivism regards the mind as a living system鈥攁n organism subject to biological and ecological demands. It emphasizes the environment in which we learn as essential. In this model, cognition is the process by which an organism understands itself and the world around it for survival.

It depends on the model

The central question is whether AI can be more ethical than humans. Cannon鈥檚 research demonstrates that this depends on the chosen model. If we adopt a cognitivist perspective, it is possible for AI to be more ethical than humans, albeit with some caveats. However, if we follow post-cognitivism, this is not feasible.

The broader significance of this conclusion lies in recognizing that existing AI research is exclusively based on cognitivist assumptions about the mind and cognition. This may mean that AI research could miss crucial aspects when modeling natural cognition

Michael Cannon defended his PhD thesis 鈥�' on Tuesday 25 June 2024. Promotors: Vincent M眉ller and Elizabeth O'Neill