Using fully autonomous AI in an inherently human activity like science and then advocating for humans in the loop puts humans in a secondary role and frames us as an add-on rather than the core. This flawed framing is against the human spirit of science."
Rather than positioning humans to verify what autonomous AI produces, it is more valuable to develop AI systems that verify and strengthen human-generated science. Instead of humans in the loop, we need AI in the loop where necessary and reliable to support and augments human judgment.
If nothing, state-level regulation could stop AI companies from taking advantage of local vulnerabilities (e.g., affordable land and energy prices) and prevent them from benefitting from regulatory loopholes, thereby reducing the risk of a race to the bottom in oversight and enforcement."
#GenAI increases the costs of over-reliance on weak & easily ‘gameable’ quantitative proxies for #research_evaluation. Brian Earp, Seb Porsdam Mann, @kristiholmes.bsky.social & I suggest that reforms should be organized around ‘distinctly human contributions' academic.oup.com/rev/article/...
The 2026 #AI index report notes that the energy consumption for only 4 prompts is comparable to charging 2 smartphones. Avoiding AI‑generated summaries reduces #environmental footprint. To search #Google WITHOUT AI Overview add "-ai" to your search query. I learned this from @Adela V. Justice.
David Resnik & I weigh options on how to define #plagiarism in light of #GenAI features. Revising definitions (to specifically include that plagiarism may also be committed when using GenAI), makes it clear that GenAI users are responsible for avoiding plagiarism. www.nature.com/articles/s42...
In today's @chicagotribune.com, I argue that 'human in the loop' is a flawed framing to think about #AI's role in #science, and that we should advocate for a slower, #human_centered model of science: www.chicagotribune.com/2026/06/04/o...
Through Advarra’s Council for Responsible Use of #AI in Clinical Trials, we worked with a diverse group from academia and industry to create a framework that introduces a practical, risk-based approach based on the degree of AI autonomy and patient impact. Access the document: hubs.la/Q04d5zKW0
Laura Caroli and I argue that the framing #AI competition as a race with #China has resulted in #deregulation and undermines the development of safer AI tools: "The real question is not who wins the race, but who builds AI systems that people can #trust. blogs.lse.ac.uk/usappblog/20...
