Can AI Technologies Make Us More (Artificially) Intelligent?
In recent discussions about artificial intelligence (AI) and human intelligence, a pressing question emerges: Can AI technologies enhance human intelligence, or do they risk making us less intelligent? Sven Nyholm’s scholarly article delves into this complex issue, exploring both the potential cognitive enhancements AI could offer and the concerns about human downgrading.
The Dual Perspectives on AI and Human Intelligence
Nyholm identifies two primary perspectives on the relationship between AI and human intelligence:
- Worry of Cognitive Decline: This viewpoint suggests that over-reliance on AI might reduce our opportunities to engage in intelligence-requiring tasks, potentially making us less intelligent over time.
- Hope for Cognitive Enhancement: Contrarily, this perspective posits that AI could serve as a cognitive enhancer, extending and enhancing our minds, allowing us to behave as if we are more intelligent.
Case Studies: AlphaGo and ChatGPT
Nyholm discusses two pivotal examples to illustrate these perspectives: AlphaGo’s victory over Lee Sedol and the capabilities of ChatGPT.
- AlphaGo and the Human Assistant: In the famous Go matches, AlphaGo, an AI program, defeated world champion Lee Sedol in four out of five games. The human assistant who executed AlphaGo’s recommended moves was able to perform at a level beyond his natural capabilities, raising the question of whether this constituted a form of human enhancement.
- ChatGPT’s Academic Potential: Similarly, Nyholm tested ChatGPT by asking it to write an essay on Martin Heidegger’s views on AI ethics. The AI produced an impressive response, arguably better than many students or even the author could produce spontaneously. This example highlights the potential of AI to enhance writing abilities.
Extending the Mind with AI
A crucial aspect of Nyholm’s argument is the “extended mind” thesis, originally proposed by Andy Clark and David Chalmers. This thesis suggests that technologies can become part of our cognitive processes, effectively extending our minds. According to this view, if AI technologies are integrated into our cognitive systems, they could enhance our cognitive abilities.
Criteria for Cognitive Enhancement
Nyholm outlines several criteria to determine when AI technologies can be considered extensions of our minds:
- Active Causal Roles: The components inside and outside of a person’s body should affect each other.
- Regulation of Behavior: The components should help regulate the behavior of the agent in ways that a mind typically does.
- Dependence on External Parts: If the external parts (AI technologies) are removed, the overall system’s capacities should diminish.
These criteria suggest that not all AI technologies qualify as cognitive enhancers. For instance, highly autonomous AI systems that operate independently from human input might not meet these criteria.
.
Cognitive Enhancement on Individual and Group Levels
Nyholm explores cognitive enhancement on both individual and group levels:
- Individuals: AI technologies can potentially enhance individual cognitive abilities by extending their minds or enabling them to behave as if they possess enhanced cognitive abilities.
- Groups: Organized groups might also benefit from AI technologies, which can enhance the cognitive capabilities of the group as a whole, leading to more effective collective decision-making and problem-solving.
Artificial Intelligence vs. Natural Intelligence
Nyholm introduces the notion of “artificial intelligence” as distinct from “natural intelligence.” While AI technologies might not increase natural intelligence, they can enable individuals to act as if they possess higher intelligence. This distinction is crucial in understanding the potential and limitations of AI as a cognitive enhancer.
Concluding Reflections
Nyholm concludes by considering the implications of AI as a cognitive enhancer. He argues that for AI to be truly considered a cognitive enhancer, it must extend our minds and enable us to justifiably take credit for our enhanced capabilities. AI technologies that merely enable us to act as if we have enhanced intelligence, without genuinely extending our cognitive capacities, should be viewed as “artificial cognitive enhancements.”
Final Thoughts
Nyholm’s article provides a nuanced analysis of the potential for AI technologies to enhance human intelligence. By examining the conditions under which AI can extend our minds and the distinction between artificial and natural intelligence, Nyholm offers a balanced perspective on the opportunities and challenges posed by AI in the realm of cognitive enhancement.
In essence, while AI technologies hold significant promise for enhancing human cognitive abilities, it is crucial to consider how these technologies are integrated into our cognitive processes and the extent to which they truly extend our minds. This thoughtful approach will help ensure that AI serves as a genuine enhancer of human intelligence rather than a crutch that diminishes our intellectual capabilities.
Note
Nyholm, S. (2024). Artificial Intelligence and Human Enhancement: Can AI Technologies Make Us More (Artificially) Intelligent? Cambridge Quarterly of Healthcare Ethics, 33(1), 76–88.