The final panel of the CCC Computing Futures Symposium, “Computing’s Human Context,” called for a paradigm shift in how we understand and develop computing technologies, urging a move towards a more interdisciplinary and human-centric approach. Moderated by Bill Regli (University of Maryland), the discussion featured insights from Sunny Consolvo (Google), Henry Farrell (Johns Hopkins University), Adam Russell (University of Southern California), and Suresh Venkatasubramanian (Brown University).
Regli kicked off the discussion by suggesting that computer science may be on the verge of a Kuhnian paradigm shift. He argued that our current models are failing to address the complexity of modern challenges in areas like artificial general intelligence (AGI), health, and robotics. The old frameworks, he proposed, are insufficient for what we are now observing, necessitating a fundamental rethinking of the discipline.
AI: A Social and Cultural Technology
A key takeaway from the conversation was the re-framing of artificial intelligence not merely as a technical tool, but as a social and cultural phenomenon. Farrell argued that Large Language Models (LLMs) are not on a path to replicate human intelligence, but rather function as powerful engines for large-scale summarization of human-generated information, akin to how markets summarize economic data or how bureaucracy categorizes social information. This perspective underscores the need for computer scientists to engage deeply with the social sciences to understand the societal implications of their creations.
When the Electron Can Think
Adam Russell echoed Farrell’s sentiment, emphasizing that Generative AI (GenAI) is unique in its ability to disrupt our cultural categories and evoke strong emotional responses. He captured this complexity by asking his colleagues to recall the Nobel-winning physicist Murray Gell-Mann’s exhortation to “imagine how much harder physics would be if electrons could think.” Russell posited that AI is a uniquely “liminal” technology – existing on a threshold between what we know and what is emerging – because it disrupts our fundamental cultural categories and actively shapes our collective fictions, forcing us to grapple with core questions of identity. He also advocated for technology developers to focus on a clear vision for the future: “It is not safety for safety’s sake or innovation for innovation’s sake.”
Navigating the “Chaos” of Generative AI
Consolvo suggested that the Cynefin Framework could be a helpful tool to use to describe the state of Generative AI, suggesting that we are probably in the “chaotic” quadrant. In this quadrant, the immediate goal shouldn’t be to find the perfect solution, but to minimize harm; it’s when novel practices can be developed. A longer-term goal would be to move the field towards the “complex” and “complicated” quadrants, where emergent and good practices can be developed, ideally getting to the “clear” quadrant, where best practices can be developed. Consolvo stressed that it’s important to understand what state (or quadrant) we’re in so that realistic expectations can be set. She also suggested the importance of interdisciplinary collaboration to navigate the space of Generative AI, and cautioned against a “move fast and break things” approach, advocating instead for a “festina lente” (or make haste slowly) approach.
The Need for Critical Evaluation and Understanding Limits
Venkatasubramanian reminded the audience that the discourse surrounding AI today often mirrors historical conversations about technology and society. He argued for a renewed focus on understanding fundamental limits of AI. Just as in other scientific fields, he contended, understanding what a system cannot do is as important as understanding what it can. This critical evaluation is crucial for developing robust and reliable systems and for providing policymakers with a realistic understanding of AI’s capabilities and shortcomings. When a problem is holistically evaluated, it might be evident that the best solution does not include technology at all. Lastly, he emphasized the need for researchers to focus on augmenting human intelligence rather than trying to replace it.
A Call for Interdisciplinary “Mind Meld”
Throughout the discussion, a strong consensus emerged on the necessity of breaking down silos between disciplines. The speakers agreed on how important it is that researchers from the field of computer science not operate in isolation (human-computer interaction, or “HCI,” was suggested as an example of a field that has embraced interdisciplinary research). The complex challenges posed by AI and other advanced computing technologies demand a “mind meld” of approaches from across computer science, the social sciences, humanities, and other fields. To create more robust, ethical, and beneficial technologies, the perspectives of communities and end users also need to be integrated.
The conversation concluded with a call to action: to foster an environment that incentivizes interdisciplinary work, to move beyond field-specific jargon and in-group mentalities, and to approach the development of technology with a deep respect for the complex human contexts in which it will be deployed. As we move forward, the message from this panel is clear: the future of computing depends on our ability to see it not just as a field of technical problems, but as a deeply human endeavor.
Wrap-Up: Further Reading
If you enjoyed this content, we encourage you to check out the Community Driven Approaches to Research in Technology & Society CCC Workshop Report, and the Catalyzing Interdisciplinary Computing Research: Best Practices for Researchers, Funders and Organizational Leadership for more recent insights about methods for interdisciplinary collaborations, and the need for considering the human context when designing, developing, and deploying computing technologies.
