Author: Chris Lenhardt
The 19 September 2024 ACTS Webinar presentation by Dr. James Spohrer raised many fascinating questions related to potential AI use in science teams. These questions ranged from the practical, “how can science teams leverage AI to facilitate both team work and task work?” to the metaphysical when thinking in terms of his ideas related to creating and using digital twins. However, science teams have been using, creating, and wrestling with the effects of technology on their work for hundreds of years. How will adding AI to science teams change how science teams conduct their work and what might be other transformative outcomes?
Clearly, generative AI tools can potentially improve science team efficiency in the context of teamwork, automating meeting note creation, summarizing relevant literature, or facilitating and tracking project resources. Less clear, but tantalizing, is the potential to use AI for task work such as facilitating shared understanding and potentially reducing conflict, or providing new insights; all suggested by Dr. Sporher.
Arguably what makes AI potentially significantly different from prior technologies is the level of agency it may ultimately bring as a member of a science team. One can speculate on this, even without considering sentient AI or whether AI may someday be accorded status on the team as an independent contributor. Given that the implications of this new technology are not fully understood we might look to other considerations of technology-science team interactions: 1) the need to include technical specialists, 2) how might AI change science team social organization, and 3) ethical implications.
Well-resourced multidisciplinary science teams increasingly include specialists of all kinds–information technology experts, project managers, communications and engagement specialists, data curators, and more. These specialists bring, leverage, or create technology as part of their role. It would not be surprising to see AI specialists added to future science teams.1 The potential effects on science team social organization include the potential for AI agents to supplant existing roles (e.g. project manager, developers, and scientists). Finally, given the many unresolved ethical issues2 related to AI, science teams may want to include an AI ethicist empowered to give voice to these concerns.
While AI bots or digital twins may not just yet be a ‘member’ of a science team, we will continue to see more uses of generative AI to accomplish both team work (think automated meeting notes) and task work (think AI covering the full science life cycle–hypothesis generation, designing and running experiments, and collecting and analyzing the results). The pressing question in the context of team science will be how does the incorporation of these technologies influence the way these teams function and the influence they may have on scientific understanding of the world around us. These should be considered carefully by any science team seeking to add AI to the team.
1Here I don’t mean technical experts like data scientists, but a technical specialist like a lab technician operating a mass spectrometer.
2Including poor quality of data inputs leading to biased results, or the potential independent agency of AI, who is responsible for AI errors, how do you acknowledge AI in papers or presentations, and intellectual property questions related to the training datasets for AI.
