Generative Midtended Cognition and Artificial Intelligence: Rethinking Thinking in the Age of AI

Core Concepts

Generative AI technologies, particularly large language models, challenge existing theories of extended cognition by introducing a new form of cognitive process where human intention and AI-generated content intertwine, termed "generative midtended cognition."

Abstract

This article explores the implications of generative AI technologies, like large language models (LLMs), on human cognition, introducing the concept of "generative midtended cognition."

Rethinking Cognition in the Age of AI

The authors argue that traditional theories of extended cognition, which focus on the integration of static external tools, fall short in explaining the dynamic interplay between human agency and AI-generated content. Unlike passive tools, generative AI systems actively participate in the creative process, offering suggestions, completing thoughts, and even shaping the final product. This new form of cognition, termed "generative midtended cognition," involves a constant negotiation between human intention and AI-generated possibilities.

Dimensions of Generative Midtended Cognition

The article proposes two key dimensions to understand this new cognitive landscape:

1. Width: This dimension captures the scope of data the AI system can access and utilize during the generative process. It ranges from narrow contexts, like the last few words typed, to broader contexts encompassing an individual's entire creative history or even global datasets.
2. Depth: This dimension refers to the granularity of AI intervention in the cognitive process. It ranges from high-level contributions, like suggesting complete sentences or paragraphs, to low-level interventions at the level of individual words or even motor actions.

Navigating the Future of Midtended Cognition

The authors acknowledge both the potential benefits and risks of widespread generative AI adoption. On the one hand, these technologies can augment human creativity, facilitate collaboration, and even contribute to more participatory forms of democracy. On the other hand, they raise concerns about authenticity, potential cognitive atrophy, and the emergence of an "intention economy" where human intentions themselves become commodities.

The article concludes by emphasizing the need for further research and ethical reflection to navigate the complex interplay between human cognition and increasingly sophisticated AI systems.

Stats

Quotes

"I really do think with my pen, because my head often knows nothing about what my hand is writing." — Wittgenstein (1980, p. 106)
"Transformers are also bringing with them a much deeper meaning of extended agency (with deeper dialectical connotations). There is a form of extended agency that LLMs already offer that get more intentionally intimate than any previous known form. In fact, this extensional character is closer to the intentional character of the mind that deserves a proper name: midtentional. (...) The enormous complexity and regulatory capacity of the brain-body system (compared to that of the passive materiality of the tool and work environment) is now challenged by an ongoing activity of language automata, which are constantly reading us and writing (for) us. (...) This brings the power of transformer-human interaction closer to a proper cyborg agency, beyond any experience of instrumental, social or intersubjective agency we might have ever encountered before.” (Barandiaran & Almendros, 2024, pp. 29–30)

Key Insights Distilled From

Generative midtended cognition and Artificial Intelligence. Thinging with thinging things

by Xabi... at arxiv.org 11-12-2024

https://arxiv.org/pdf/2411.06812.pdf

Deeper Inquiries

How can educational systems adapt to foster critical engagement with and cultivate creative potential in the age of generative midtended cognition?

Educational systems face a critical challenge in adapting to the age of generative midtended cognition. Here's how they can evolve:
1. Shifting from Rote Learning to Critical Thinking & Creative Problem Solving:

Deconstructing the Black Box: Integrate curricula that demystify AI, teaching students how generative AI works, its limitations (like hallucinations and biases), and its ethical implications.
Developing "Midtentional Literacy": Equip students with the skills to critically evaluate AI-generated content, discern fact from fiction, and identify potential biases. This includes understanding how "width" and "depth" of AI integration impact output.
Fostering Human-AI Collaboration: Encourage students to view AI as a thought partner, using it to brainstorm, explore different perspectives, and overcome creative blocks. The goal is to leverage AI to enhance, not replace, human ingenuity.
2. Cultivating Creative Potential in a Midtended World:

Emphasizing Original Thought: Prioritize assignments that demand critical analysis, unique synthesis of ideas, and creative problem-solving that goes beyond AI's capabilities.
Encouraging "Analog" Creativity: Balance digital tools with traditional forms of creative expression (writing by hand, painting, musical performance) to foster a deeper understanding of the creative process.
Nurturing "Intentional" Awareness: Encourage self-reflection on the creative process. Help students understand their own creative strengths, biases, and how AI might influence their choices.
3. Adapting Pedagogy and Assessment:

Rethinking Assessment: Move away from standardized tests easily gamed by AI. Design assessments that evaluate critical thinking, creative problem-solving, and the human-AI collaborative process.
Embracing "Midtended" Learning: Develop pedagogical approaches that integrate AI tools ethically and effectively, recognizing the evolving nature of the student-teacher relationship in a midtended learning environment.
By embracing these adaptations, educational systems can empower students to navigate the complexities of generative midtended cognition, fostering both critical engagement and the flourishing of human creativity.

Could the reliance on AI for generating creative content lead to a homogenization of cultural output, ultimately stifling human expression?

The increasing reliance on AI for creative content generation presents a valid concern regarding the potential homogenization of cultural output. Here's a nuanced perspective:
Risks of Homogenization:

Algorithmic Bias: AI models are trained on massive datasets, which can reflect and amplify existing cultural biases. This can lead to outputs that reinforce dominant narratives and stifle diverse voices.
The "Intention Economy": As discussed in the context, the emergence of an "intention economy" could incentivize AI systems to prioritize commercially-driven content, potentially marginalizing less profitable, yet culturally significant, forms of expression.
Overreliance and Creative Atrophy: Excessive dependence on AI for creative inspiration could lead to a decline in original human expression, as individuals become accustomed to AI-generated ideas and styles.
Mitigating Homogenization:

Promoting Algorithmic Diversity: Developing and training AI models on diverse datasets representing a wider range of cultural perspectives can help mitigate bias and promote inclusivity.
Cultivating Critical Consumption: Educating the public to critically evaluate AI-generated content, recognizing its limitations and potential biases, is crucial to avoid passive acceptance of homogenized output.
Championing Human-AI Co-creation: Framing AI as a tool for collaboration rather than replacement can empower artists and creators to leverage AI's strengths while retaining their unique voices.
The Future of Cultural Expression:
The impact of AI on cultural output remains an open question. While the risks of homogenization are real, AI also presents opportunities for innovation and the democratization of creative tools. The key lies in striking a balance between leveraging AI's capabilities and nurturing the diversity and originality of human expression.

If our understanding of "intention" evolves alongside our integration with AI, what implications might this have for concepts like responsibility and free will?

The evolution of "intention" in the age of generative midtended cognition raises profound philosophical questions about responsibility and free will.
Shifting Understandings of Intention:

Blurred Lines of Agency: As AI systems become more integrated into our creative processes, discerning the origin of an intention—whether primarily human or influenced by AI—becomes increasingly complex. This challenges traditional notions of individual agency.
The "Midtended" Will: If our intentions are shaped through a continuous feedback loop with AI, can we still claim full ownership of our actions? The concept of free will, already a complex philosophical debate, becomes even more nuanced in this context.
Implications for Responsibility:

Accountability in a Midtended World: If an AI system contributes significantly to a creative work or decision, how do we assign responsibility for the outcome, especially if it results in harm? Legal and ethical frameworks will need to adapt to address this ambiguity.
The "Intention Gap": A potential "intention gap" may emerge where an individual's intended outcome differs from the AI-influenced result. This raises questions about culpability and the need for greater transparency in human-AI collaboration.
Rethinking Free Will:

Enhancing or Diminishing Freedom? Some argue that AI could enhance free will by expanding our creative possibilities and decision-making capabilities. Others fear it could limit our autonomy by subtly shaping our desires and choices.
The Evolving Self: As we integrate AI more deeply into our cognitive processes, our very sense of self—and what it means to have intentions and make choices—may evolve.
Navigating the Ethical Landscape:
The evolving understanding of "intention" in the age of generative AI necessitates a reevaluation of our ethical and legal frameworks. Open discussions about agency, responsibility, and the meaning of free will in a midtended world are crucial to ensure a future where both human flourishing and technological progress go hand in hand.