Upcoming Event: CosmicAI Seminar

1) Responsible AI for Science: Assessing Potential Benefits vs. Risks 2) Who Own Creativity and Who Does the Work? Trade-offs in LLM-Supported Research Ideation

1) Matt Lease 2) Houjiang Liu,

Abstract

**CSEM students can receive credit for attending CosmicAI seminars**

Tacos will be served.

Part 1 Abstract: How can we best harness modern AI capabilities to accelerate scientific discovery? Any change from the status quo can bring about both beneficial and adverse impacts, and we have seen many recent examples of new AI technologies causing societal harm as well as good. For this reason, responsible AI governance advocates for advance assessment of both potential risks and benefits that could stem from future adoption of AI technologies. Rather than react to such harms after they have already occurred, it is far better to anticipate, prevent, and mitigate such harms when possible. Moreover, we should convince ourselves that potential benefits outweigh potential risks before pursuing AI adoption. This parallels how institutional review boards (IRBs) similarly require proposed human subject research studies to both assess the level of risk posed and to justify potential benefits outweighing risks prior to study approval. My talk will briefly introduce foundations of responsible AI governance before reflecting on some of the potential benefits vs. risks posed by greater adoption of AI in scientific practice. The talk will further serve to help bridge ongoing work in UT Austin’s campus-wide responsible AI initiative, Good Systems, with that of the NSF-Simons AI Institute for Cosmic Origins, CosmicAI, led by UT Austin.  

Part 2 Abstract: LLM-based agents offer new potential to accelerate science and reshape research work. However, the quality of final intellectual outcomes can vary significantly depending on degrees of human involvement. How can we best use these tools to augment scientific creativity without undermining aspects of contribution and ownership that drive research? This work investigates varying levels of human controls (minimum, medium, and intensive) over an agentic research ideation workflow, revealing how they affect AI creative support and the researcher personal effort they invest. Our mixed-methods study with 54 researchers suggests three findings: 1) the perceived creativity support of AI does not simply increase linearly with greater control; 2) human effort remains consistent across control levels but the nature of work shifts from ideating to verifying; and 3) ownership of the final intellectual work becomes a negotiated outcome between human and AI. Future tool design for AI-driven automated research ideation should empower researchers, allowing them to experience a greater sense of ownership over more powerful ideas, rather than reducing them to operators of a machine that controls the creative process.

Biography

Matthew Lease is a Professor of Information at UT Austin, a Distinguished Member of the Association for Computing Machinery (ACM), and a Senior Member of the Association for the Advancement of Artificial Intelligence (AAAI). He also serves as CosmicAI Co-Director and is a faculty founder and leader of Good Systems. Recent honors include the 2024 AAAI HCOMP Test of Time Award, the 2024 IEEE/ACM ASE Most Influential Paper Award, and a ACM CSCW 2024 Honorable Mention. Beyond scholarship, Lease has also informed responsible AI policy at state and national levels. 

Houjiang Liu is a PhD student studying human-centered computing at the school of information. His recent research focuses on examining how LLM-based agents shape different stages of research activities.