MIT FutureTech

The MIT FutureTech team led by Neil Thompson brought together leaders from academia, industry, and government to explore AI’s evolving role in scientific discovery. As a research affiliate, I focus on measuring how innovations in emerging technologies drive productivity gains and shape distributional impacts. This event offered valuable insights into these dynamics—here’s a look at the key takeaways. ?? ???????????????????????? ???????????????????? ?????????????????? AI is advancing scientific research by enabling faster, deeper insights—unlocking breakthroughs in drug discovery, diagnostics, material science, economic modeling and more.?This tech-driven expansion allows scientists to tackle increasingly complex questions at an unprecedented scale. Danial Lashkari Regina Barzilay ROSS King Lilach M. ?? ???????????????????? ???????? ?????? ???????????????????? ???????????? AI’s adaptive modeling capabilities are revolutionizing how researchers interpret data, enabling real-time hypothesis refinement and theory testing. With machine learning at the core, science is shifting towards a more dynamic, responsive research paradigm. Yuji Roh Boris Kozinsky Manuela Veloso Stephen Wolfram ?? ?????????? ????????????????: ????-?????????????????????????????? ?????? ????’?? ???????????????????????????? The rise of AI is accompanied by increasing centralization, with a few players controlling critical infrastructure, capabilities, and models. This shift raises key questions about who benefits from AI advancements and how limited access could impact innovation in the scientific community within and across countries. Jan Eeckhout Rada Mihalcea Varun Chandola Haydn Belfield ?? ?????????????? ?????? ?????????????????? ???????????????????? ???? ????-???????????? ?????????????? AI’s integration into research comes with risks—bias in models, privacy concerns, and the need for rigorous content verification. Strong ethical frameworks are essential to balance responsible innovation with scientific integrity and public trust. Lei Li Dashun Wang Keyon Vafa Eamon Duede MIT Computer Science and Artificial Intelligence Laboratory (CSAIL)Massachusetts Institute of Technology MIT Technology Review

At a recent lab meeting, Ekaterina Prytkova and Simone Vannuccini from University of Sussex and Université C?te d'Azur presented their paper 'On the basis of brain: neural-network-inspired changes in general-purpose chips".

Laurence Ales presented "How It’s Made: A General Theory of the Labor Implications of Technology Change" at a recent lab meeting. We develop a general theory relating technology change and skill demand. Performers (human or machine) face stochastic issues that must be solved in order to complete tasks. Firms choose how production tasks are divided into steps, the rate at which steps need to be completed, and the skill of the performer assigned to a step. Longer steps are more complex. Performers face a tradeoff between the complexity of their step and the rate at which they can perform. Human performers tend to have an advantage in complex steps while machine performers have an advantage in high rates. The cost of fragmenting tasks into steps and the cost of allocating performers to multiple steps are both central to the theory. We derive the optimal division of tasks, the level of automation, and the demand for workers of different skill levels. The theory predicts that automation generates skill polarization at lower production volumes and is upskilling at higher volumes; in addition, the theory implies that a reduction in fragmentation costs (such as interchangeable parts) increases the demand for low skill; and that technology change that raises the cost of fragmenting tasks (such as parts consolidation) reduces the dispersion of skill demand. We find counterparts to the theory across a range of contexts and time periods, including the Hand-Machine Labor Study covering mechanization and process improvement at the end of the 19th century and in contemporary automotive body assembly and optoelectronic semiconductor manufacturing"

Neil Thompson recently delivered a keynote outlining the AI Risk Repository at the ALL IN AI conference in Canada. The session was organized in partnership with the Coalition pour la diversité des expressions culturelles / for the diversity of cultural expressions and supported by the Ministère de la Culture. ALL IN is an initiative of Scale AI and co-organized with CEIMIA and Mila - Quebec Artificial Intelligence Institute. Thanks to Marie-Julie Desrochers and Orélie Br?let for their invitation and support. https://lnkd.in/ejdKaDkA #ArtificialIntelligence #Technology

"Exploiting a major internet control shock in 2014, I find that Chinese data-intensive firms gained 26 percent in revenue over other Chinese firms as the result of internet control. The same shock incurred a 10 percent decline in research quality from Chinese researchers, conditional on the knowledge intensity of their discipline. It also reduced the research quality from Chinese researchers relative to their US counterparts by 22 percent in all disciplines" New research from our Affiliate Professor, Meicen Sun.

At a recent lab meeting, Frank Nagle presented his paper 'Generative AI and the Nature of Work': "We seek to build upon research on #AI and productivity to better understand how #GenAI changes how people do work. We look at how the release of a GenAI coding tool (GitHub Copilot) changed how developers allocate their efforts to different types of tasks. We find that GenAI leads workers to spend more time on core work activities and less time on managerial tasks. We show two mechanisms drive this effect - workers with GenAI allocate more of their work efforts to things they can do by themselves (and less to collaborative work) and also do more exploration (new projects, new languages, etc.) and less exploitation (existing projects). Further we find the effects are greater for workers with lower ability. Finally, we do a back-of-the-envelope calculation and show that using GenAI allows developers to start coding in languages that have higher wages, leading to a labor market value impact of nearly $500 million (this would likely diminish in the long run). Though our empirical setting is open source software #OSS, we argue, and find evidence, that the results generalize to private work settings as well."

How are big trends in computing shaping science? The Computing Community Consortium (CCC) discuss a panel featuring Gabriel and Jayson, and moderated by Neil at the recent American Association for the Advancement of Science (AAS) conference. https://lnkd.in/eNXFRyRT

Some photos from our recent workshop exploring the role of AI in science!

Today is the second day of our workshop on the role of AI in science! First up, we have Neil Thompson introducing session 3: De-democratization and concentration of power in AI. See the full agenda here: https://lnkd.in/e_nzMDva

Today is the first day of the MIT FutureTech workshop on the role of AI in science! We will be posting updates throughout the day. We are bringing together some of the most influential voices in AI and scientific research to discuss the profound ways artificial intelligence is reshaping science as we know it. Topics include: - The role of AI in advancing scientific research - AI in drug discovery and medical breakthroughs - The ethical implications and societal impacts of AI - Challenges and future directions for AI-driven science See full agenda here: https://lnkd.in/e_nzMDva MIT Sloan School of Management CSAIL MIT MIT Initiative on the Digital Economy Massachusetts Institute of Technology