Paper by Joshua S. Gans: “This paper examines how the introduction of artificial intelligence (AI), particularly generative and large language models capable of interpolating precisely between known data points, reshapes scientists’ incentives for pursuing novel versus incremental research. Extending the theoretical framework of Carnehl and Schneider (2025), we analyse how decision-makers leverage AI to improve precision within well-defined knowledge domains. We identify conditions under which the availability of AI tools encourages scientists to choose more socially valuable, highly novel research projects, contrasting sharply with traditional patterns of incremental knowledge growth. Our model demonstrates a critical complementarity: scientists strategically align their research novelty choices to maximise the domain where AI can reliably inform decision-making. This dynamic fundamentally transforms the evolution of scientific knowledge, leading either to systematic “stepping stone” expansions or endogenous research cycles of strategic knowledge deepening. We discuss the broader implications for science policy, highlighting how sufficiently capable AI tools could mitigate traditional inefficiencies in scientific innovation, aligning private research incentives closely with the social optimum…(More)”.
The Age of AI in the Life Sciences: Benefits and Biosecurity Considerations
Report by the National Academies of Sciences, Engineering, and Medicine: “Artificial intelligence (AI) applications in the life sciences have the potential to enable advances in biological discovery and design at a faster pace and efficiency than is possible with classical experimental approaches alone. At the same time, AI-enabled biological tools developed for beneficial applications could potentially be misused for harmful purposes. Although the creation of biological weapons is not a new concept or risk, the potential for AI-enabled biological tools to affect this risk has raised concerns during the past decade.
This report, as requested by the Department of Defense, assesses how AI-enabled biological tools could uniquely impact biosecurity risk, and how advancements in such tools could also be used to mitigate these risks. The Age of AI in the Life Sciences reviews the capabilities of AI-enabled biological tools and can be used in conjunction with the 2018 National Academies report, Biodefense in the Age of Synthetic Biology, which sets out a framework for identifying the different risk factors associated with synthetic biology capabilities…(More)”
Generative AI in Transportation Planning: A Survey
Paper by Longchao Da: “The integration of generative artificial intelligence (GenAI) into transportation planning has the potential to revolutionize tasks such as demand forecasting, infrastructure design, policy evaluation, and traffic simulation. However, there is a critical need for a systematic framework to guide the adoption of GenAI in this interdisciplinary domain. In this survey, we, a multidisciplinary team of researchers spanning computer science and transportation engineering, present the first comprehensive framework for leveraging GenAI in transportation planning. Specifically, we introduce a new taxonomy that categorizes existing applications and methodologies into two perspectives: transportation planning tasks and computational techniques. From the transportation planning perspective, we examine the role of GenAI in automating descriptive, predictive, generative simulation, and explainable tasks to enhance mobility systems. From the computational perspective, we detail advancements in data preparation, domain-specific fine-tuning, and inference strategies such as retrieval-augmented generation and zero-shot learning tailored to transportation applications. Additionally, we address critical challenges, including data scarcity, explainability, bias mitigation, and the development of domain-specific evaluation frameworks that align with transportation goals like sustainability, equity, and system efficiency. This survey aims to bridge the gap between traditional transportation planning methodologies and modern AI techniques, fostering collaboration and innovation. By addressing these challenges and opportunities, we seek to inspire future research that ensures ethical, equitable, and impactful use of generative AI in transportation planning…(More)”.
Launch: A Blueprint to Unlock New Data Commons for Artificial Intelligence (AI)
Blueprint by Hannah Chafetz, Andrew J. Zahuranec, and Stefaan Verhulst: “In today’s rapidly evolving AI landscape, it is critical to broaden access to diverse and high-quality data to ensure that AI applications can serve all communities equitably. Yet, we are on the brink of a potential “data winter,” where valuable data assets that could drive public good are increasingly locked away or inaccessible.
Data commons — collaboratively governed ecosystems that enable responsible sharing of diverse datasets across sectors — offer a promising solution. By pooling data under clear standards and shared governance, data commons can unlock the potential of AI for public benefit while ensuring that its development reflects the diversity of experiences and needs across society.
To accelerate the creation of data commons, The Open Data Policy, today, releases “A Blueprint to Unlock New Data Commons for AI” — a guide on how to steward data to create data commons that enable public-interest AI use cases…the document is aimed at supporting libraries, universities, research centers, and other data holders (e.g. governments and nonprofits) through four modules:
- Mapping the Demand and Supply: Understanding why AI systems need data, what data can be made available to train, adapt, or augment AI, and what a viable data commons prototype might look like that incorporates stakeholder needs and values;
- Unlocking Participatory Governance: Co-designing key aspects of the data commons with key stakeholders and documenting these aspects within a formal agreement;
- Building the Commons: Establishing the data commons from a practical perspective and ensure all stakeholders are incentivized to implement it; and
- Assessing and Iterating: Evaluating how the commons is working and iterating as needed.
These modules are further supported by two supplementary taxonomies. “The Taxonomy of Data Types” provides a list of data types that can be valuable for public-interest generative AI use cases. The “Taxonomy of Use Cases” outlines public-interest generative AI applications that can be developed using a data commons approach, along with possible outcomes and stakeholders involved.
A separate set of worksheets can be used to further guide organizations in deploying these tools…(More)”.
Funding the Future: Grantmakers Strategies in AI Investment
Report by Project Evident: “…looks at how philanthropic funders are approaching requests to fund the use of AI… there was common recognition of AI’s importance and the tension between the need to learn more and to act quickly to meet the pace of innovation, adoption, and use of AI tools.
This research builds on the work of a February 2024 Project Evident and Stanford Institute for Human-Centered Artificial Intelligence working paper, Inspiring Action: Identifying the Social Sector AI Opportunity Gap. That paper reported that more practitioners than funders (by over a third) claimed their organization utilized AI.
“From our earlier research, as well as in conversations with funders and nonprofits, it’s clear there’s a mismatch in the understanding and desire for AI tools and the funding of AI tools,” said Sarah Di Troia, Managing Director of Project Evident’s OutcomesAI practice and author of the report. “Grantmakers have an opportunity to quickly upskill their understanding – to help nonprofits improve their efficiency and impact, of course, but especially to shape the role of AI in civil society.”
The report offers a number of recommendations to the philanthropic sector. For example, funders and practitioners should ensure that community voice is included in the implementation of new AI initiatives to build trust and help reduce bias. Grantmakers should consider funding that allows for flexibility and innovation so that the social and education sectors can experiment with approaches. Most importantly, funders should increase their capacity and confidence in assessing AI implementation requests along both technical and ethical criteria…(More)”.
Bridging the Data Provenance Gap Across Text, Speech and Video
Paper by Shayne Longpre et al: “Progress in AI is driven largely by the scale and quality of training data. Despite this, there is a deficit of empirical analysis examining the attributes of well-established datasets beyond text. In this work we conduct the largest and first-of-its-kind longitudinal audit across modalities–popular text, speech, and video datasets–from their detailed sourcing trends and use restrictions to their geographical and linguistic representation. Our manual analysis covers nearly 4000 public datasets between 1990-2024, spanning 608 languages, 798 sources, 659 organizations, and 67 countries. We find that multimodal machine learning applications have overwhelmingly turned to web-crawled, synthetic, and social media platforms, such as YouTube, for their training sets, eclipsing all other sources since 2019. Secondly, tracing the chain of dataset derivations we find that while less than 33% of datasets are restrictively licensed, over 80% of the source content in widely-used text, speech, and video datasets, carry non-commercial restrictions. Finally, counter to the rising number of languages and geographies represented in public AI training datasets, our audit demonstrates measures of relative geographical and multilingual representation have failed to significantly improve their coverage since 2013. We believe the breadth of our audit enables us to empirically examine trends in data sourcing, restrictions, and Western-centricity at an ecosystem-level, and that visibility into these questions are essential to progress in responsible AI. As a contribution to ongoing improvements in dataset transparency and responsible use, we release our entire multimodal audit, allowing practitioners to trace data provenance across text, speech, and video…(More)”.
Artificial intelligence for modelling infectious disease epidemics
Paper by Moritz U. G. Kraemer et al: “Infectious disease threats to individual and public health are numerous, varied and frequently unexpected. Artificial intelligence (AI) and related technologies, which are already supporting human decision making in economics, medicine and social science, have the potential to transform the scope and power of infectious disease epidemiology. Here we consider the application to infectious disease modelling of AI systems that combine machine learning, computational statistics, information retrieval and data science. We first outline how recent advances in AI can accelerate breakthroughs in answering key epidemiological questions and we discuss specific AI methods that can be applied to routinely collected infectious disease surveillance data. Second, we elaborate on the social context of AI for infectious disease epidemiology, including issues such as explainability, safety, accountability and ethics. Finally, we summarize some limitations of AI applications in this field and provide recommendations for how infectious disease epidemiology can harness most effectively current and future developments in AI…(More)”.
Future of AI Research
Report by the Association for the Advancement of Artificial Intelligence: “As AI capabilities evolve rapidly, AI research is also undergoing a fast and significant transformation along many dimensions, including its topics, its methods, the research community, and the working environment. Topics such as AI reasoning and agentic AI have been studied for decades but now have an expanded scope in light of current AI capabilities and limitations. AI ethics and safety, AI for social good, and sustainable AI have become central themes in all major AI conferences. Moreover, research on AI algorithms and software systems is becoming increasingly tied to substantial amounts of dedicated AI hardware, notably GPUs, which leads to AI architecture co-creation, in a way that is more prominent now than over the last 3 decades. Related to this shift, more and more AI researchers work in corporate environments, where the necessary hardware and other resources are more easily available, compared to academia, questioning the roles of academic AI research, student retention, and faculty recruiting. The pervasive use of AI in our daily lives and its impact on people, society, and the environment makes AI a socio-technical field of study, thus highlighting the need for AI researchers to work with experts from other disciplines, such as psychologists, sociologists, philosophers, and economists. The growing focus on emergent AI behaviors rather than on designed and validated properties of AI systems renders principled empirical evaluation more important than ever. Hence the need arises for well-designed benchmarks, test methodologies, and sound processes to infer conclusions from the results of computational experiments. The exponentially increasing quantity of AI research publications and the speed of AI innovation are testing the resilience of the peer-review system, with the immediate release of papers without peer-review evaluation having become widely accepted across many areas of AI research. Legacy and social media increasingly cover AI research advancements, often with contradictory statements that confuse the readers and blur the line between reality and perception of AI capabilities. All this is happening in a geo-political environment, in which companies and countries compete fiercely and globally to lead the AI race. This rivalry may impact access to research results and infrastructure as well as global governance efforts, underscoring the need for international cooperation in AI research and innovation.
In this overwhelming multi-dimensional and very dynamic scenario, it is important to be able to clearly identify the trajectory of AI research in a structured way. Such an effort can define the current trends and the research challenges still ahead of us to make AI more capable and reliable, so we can safely use it in mundane but also, most importantly, in high-stake scenarios.
This study aims to do this by including 17 topics related to AI research, covering most of the transformations mentioned above. Each chapter of the study is devoted to one of these topics, sketching its history, current trends and open challenges…(More)”.
AI could supercharge human collective intelligence in everything from disaster relief to medical research
Article by Hao Cui and Taha Yasseri: “Imagine a large city recovering from a devastating hurricane. Roads are flooded, the power is down, and local authorities are overwhelmed. Emergency responders are doing their best, but the chaos is massive.
AI-controlled drones survey the damage from above, while intelligent systems process satellite images and data from sensors on the ground and air to identify which neighbourhoods are most vulnerable.
Meanwhile, AI-equipped robots are deployed to deliver food, water and medical supplies into areas that human responders can’t reach. Emergency teams, guided and coordinated by AI and the insights it produces, are able to prioritise their efforts, sending rescue squads where they’re needed most.
This is no longer the realm of science fiction. In a recent paper published in the journal Patterns, we argue that it’s an emerging and inevitable reality.
Collective intelligence is the shared intelligence of a group or groups of people working together. Different groups of people with diverse skills, such as firefighters and drone operators, for instance, work together to generate better ideas and solutions. AI can enhance this human collective intelligence, and transform how we approach large-scale crises. It’s a form of what’s called hybrid collective intelligence.
Instead of simply relying on human intuition or traditional tools, experts can use AI to process vast amounts of data, identify patterns and make predictions. By enhancing human decision-making, AI systems offer faster and more accurate insights – whether in medical research, disaster response, or environmental protection.
AI can do this, by for example, processing large datasets and uncovering insights that would take much longer for humans to identify. AI can also get involved in physical tasks. In manufacturing, AI-powered robots can automate assembly lines, helping improve efficiency and reduce downtime.
Equally crucial is information exchange, where AI enhances the flow of information, helping human teams coordinate more effectively and make data-driven decisions faster. Finally, AI can act as social catalysts to facilitate more effective collaboration within human teams or even help build hybrid teams of humans and machines working alongside one another…(More)”.
Governing in the Age of AI: Building Britain’s National Data Library
Report by the Tony Blair Institute for Global Change: “The United Kingdom should lead the world in artificial-intelligence-driven innovation, research and data-enabled public services. It has the data, the institutions and the expertise to set the global standard. But without the right infrastructure, these advantages are being wasted.
The UK’s data infrastructure, like that of every nation, is built around outdated assumptions about how data create value. It is fragmented and unfit for purpose. Public-sector data are locked in silos, access is slow and inconsistent, and there is no system to connect and use these data effectively, or any framework for deciding what additional data would be most valuable to collect given AI’s capabilities.
As a result, research is stalled, AI adoption is held back, and the government struggles to plan services, target support and respond to emerging challenges. This affects everything from developing new treatments to improving transport, tackling crime and ensuring economic policies help those who need them. While some countries are making progress in treating existing data as strategic assets, none have truly reimagined data infrastructure for an AI-enabled future…(More)”