AI in RA
« AI in RA

#5 — US government $5B AI, White House redirects research funds

July 22, 2026
The US government is dedicating $5 billion to an initiative focused on using artificial intelligence to accelerate scientific research, including drug discovery. The program involves 15 federal agencies and is supported by $40 million in AI computing credits from Microsoft. Other developments include a White House plan to redirect federal research funding from universities to AI companies and the use of an AI tool that identified a large volume of potentially fraudulent cancer research papers.

Businesses increasingly deploy AI for workflow automation, data analysis, and task management. Key AI productivity trends include AI agent systems, multimodal AI, and AI integration into workplace software.

Generative AI enables synthetic relationships, potentially addressing loneliness but raising psychological, ethical, and societal questions. These relationships offer an alternative to existing loneliness interventions while posing risks like emotional over-reliance, distorted social expectations, and privacy concerns. Widespread adoption may reshape human relationships, requiring a research agenda to address ethical considerations; they serve as social interventions when complementing human interaction.

Studies focused on AI in agriculture include developing hybrid techniques for crop yield prediction, combining multinomial logistic regression and Yeo-Johnson transformers for crop recommendation, and applying deep neural networks for future crop market price prediction.

A research project aims to develop regional diffuse solar radiation models. Qualitative analysis investigates environmental factors influencing the clearness index-diffuse ratio relationship. Quantitative assessment develops and recommends optimal region-wise hourly solar radiation models. Daily diffuse radiation models are developed using daily environmental parameters and validated against other sites, acknowledging that measurement uncertainty impacts model accuracy.

Researchers introduced the Bias Evaluations Across Domains (BEADs) dataset, which incorporates a gold-standard annotation scheme. Experiments using this dataset revealed biases in state-of-the-art models.

In higher education, increased use of analytical and predictive tools correlates with reduced academic losses and higher program completion rates. Implementing predictive analytics and early warning systems supports timely managerial interventions and enhances student success. AI adoption facilitates a shift from reactive governance to system-based educational modeling.

A research paper develops a framework for understanding systemic risk in AI and platform governance. It critiques the EU AI Act and Digital Services Act for narrow systemic risk characterizations, identifying risks overlooked by these acts. The paper proposes reforms for systemic risk assessments and regulatory coordination.

Researchers developed a taxonomy of AI regulation capture mechanisms, identifying "Discourse & Epistemic Influence" and "Elusion of law" as recurring categories. Narratives rationalizing AI regulation capture frequently include "Regulation stifles innovation," "Red tape," and "National Interest."

Sources

  1. US Turns To AI To Tackle Long-Standing Problems, Pledges $5 Bn For Scientific Research | Times Now
    The US is investing $5 billion in AI to accelerate drug discovery, address chronic diseases, and advance scientific research by utilizing supercomputers and government data. This initiative, involving 15 federal agencies, aims to combine government datasets in areas like healthcare and critical minerals with advanced AI systems to identify patterns, generate predictions, and answer scientific questions more efficiently. Microsoft will support the program by providing $40 million worth of AI computing credits to researchers over the next three years.
  2. White House Plans to Redirect Billions in Federal Research Funding ...
    According to reports, the White House plans to redirect billions in federal research funding from universities to the development of AI. This adjustment prioritizes support for AI infrastructure and application research, reducing allocations to traditional academic institutions. Federal funds will flow to AI companies and dedicated projects, benefiting tech firms and AI startups, while university-based research faces budget pressures, accelerating the reallocation of resources between academia and industry.
  3. AI flags more than 250,000 suspicious cancer research papers ...
    A powerful new AI tool has uncovered what could be one of the biggest integrity problems in modern science. After analyzing 2.6 million cancer research papers published between 1999 and 2024, researchers identified more than 250,000 studies with writing patterns resembling papers suspected of being produced by fraudulent "paper mills."
  4. A Practical Guide for AI Governance and Higher Education - Broadfield
    For university leaders, the academic year will arrive against a backdrop of extraordinary urgency for governing artificial intelligence on campus. Engagement with unapproved AI tools threatens to create data flows that bypass institutional security controls and potentially violate federal privacy mandates. These rapid changes in the political and technological headwinds signal a clear mandate: the question is no longer whether AI governance is necessary, but whether your institution's governance infrastructure is adequate.
  5. A framework for developing university policies on generative AI governance: a cross-national comparative study
    As generative AI (GAI) becomes increasingly embedded in higher education, universities worldwide are developing policies to govern its ethical, pedagogical, and institutional use. However, these policies vary across national and institutional contexts. We undertake a cross-national analysis of GAI guidelines issued by leading universities in the United States, Japan, and China, identifying key policy orientations and proposing a structured framework to support policy development. Using an extended Technology Acceptance Model as an analytical lens, we examine five domains – Perceived…
  6. Top 10 AI productivity trends shaping work in 2026 - Businessday NG
    Businesses are increasingly deploying AI to automate workflows, analyze data, and manage tasks, transforming AI from a mere assistant into a digital coworker by 2026. This shift necessitates organizations to combine AI capabilities with strong governance and workforce training to achieve productivity gains. The article details 10 key AI productivity trends, such as the rise of AI agent systems, multimodal AI, and AI integrated into workplace software, that are reshaping professional work.
  7. White House will gut university research funding to bankroll AI push
    Key takeaways: White House plans to reroute federal research money from universities toward individual scientists and AI Critics say the shift could damage universities and inject unreliable AI into delicate R&D Administration argues AI-focused funding will accelerate discovery and strengthen America against Chinese competition
  8. Gemini and Gemini Notebook added to portfolio of AI services | Penn State University
    Penn State has integrated Google's Gemini and Gemini Notebook (formerly NotebookLM) into its AI services portfolio. Gemini is a general-purpose AI assistant for various tasks, offering expanded capabilities, multimodal intelligence, and Google Workspace integration. Gemini Notebook is a specialized, source-grounded AI research assistant that provides accurate, cited responses from uploaded documents and features like audio overviews.
  9. International Enrollment, AI Governance, and Graduate Aid Move Into Execution
    The Ecosystem Weekly: DHS visa reforms, graduate aid implementation, AI governance, and new research partnership models reshape institutional priorities.

Also this week

Full transcript
The U.S. government is committing billions to accelerate scientific research using artificial intelligence. That is one of the developments we are tracking on AI in RA, a show that follows AI's integration into academic and corporate environments. We begin with the federal government's new initiative. The U.S. government is launching a five-billion-dollar initiative to apply artificial intelligence to scientific research. Right, and the goal is to use government datasets with AI systems for things like accelerating drug discovery. It involves 15 federal agencies, and Microsoft is also contributing, offering forty million dollars in computing credits to researchers. But there's a significant change in how the funding works. The White House is reportedly planning to reallocate billions in federal research funding *away* from universities. And redirecting it toward AI infrastructure and application research. That's a move with critics, who are warning it could harm universities and also lead to unreliable AI being used in R&D. The administration's position is that this focus will speed up discovery and also bolster the country's competitive stance against China. So, while that funding debate is happening, AI is already being used to check the integrity of existing science. One tool analyzed 2.6 million cancer research papers and flagged over 250,000 studies. Flagged them for what, exactly? For writing patterns that were consistent with fraudulent 'paper mills'. So it's being used as a kind of research watchdog. This brings up the challenges of using AI inside academic institutions. Using unapproved AI tools can open up data flows that get around a university's security, which could violate privacy laws and put research data at risk. Which is why universities are trying to figure out their own guidelines. A recent cross-national analysis looked at policies from universities in the U.S., Japan, and China and found they have very different priorities. How so? U.S. guidelines tend to focus on faculty autonomy and making policies adaptable. Japanese universities put more emphasis on ethics and risk management. And Chinese universities are focused on technology application. So, based on those different approaches, the researchers developed a framework to help universities navigate this, a way to balance innovation and risk. And we're already seeing specific adoptions, like Penn State integrating Google's Gemini into its services. This idea of AI governance is becoming a central factor for all kinds of organizations, not just universities. To get any productivity gains from AI, they first have to set up governance frameworks and provide training. And it's moving into so many different areas. We're seeing businesses use AI for workflow automation and data analysis, with trends toward AI agent systems and multimodal AI. There's also the rise of what are being called synthetic relationships. Generative AI is enabling connections that could help with loneliness, but it also opens up a whole set of psychological and ethical questions. Right, prompting calls for a research agenda to study the risks, like emotional over-reliance on these systems. It's also being applied in agriculture, with hybrid techniques for predicting crop yield and neural networks for market price prediction. Other work is trying to develop regional solar radiation models. And as all these applications expand, so does the work on measuring bias. A new dataset, Bias Evaluations Across Domains, has already uncovered biases in current models. It all circles back to policy and systems. In higher education, analytical tools are being linked to higher program completion rates. And in policy analysis, new research is critiquing the EU AI Act for having too narrow a view of systemic risk. That is the program for today. We will have more updates next week. Until next time, on AI in RA.

From our community

Builders Group

A peer working group for research administrators who are actively building AI into their offices. Some members work alongside IT teams; others are piecing it together on their own. What unites the group is the work of implementation itself:

RSVP →

From our community

Join AI in Research Administration

Join the community →