X

4th AI Symposium Program

Education Fourth AI Symposium 4th AI Symposium Program

Session Descriptions

8:00 - 9:00am Keynote Session

Location: 103AB, Floor 1
Presenter: Dr. Johnny MacLean, Montana Tech University

In this keynote, Montana Tech Chancellor Johnny MacLean describes how Artificial Intelligence represents a watershed moment that will fundamentally and permanently reshape higher education, federal research priorities, and national strategy. Drawing on Montana Tech’s 125-year history of responding to past watershed moments, from electrification and world wars to energy security, environmental reckoning, and public health, Dr. MacLean frames AI not as a productivity tool, but as the central national security imperative of our time. Just as 9/11 and COVID-19 redefined global systems, AI will redefine society on a far greater scale, leaving the world unrecognizable in five to ten years. Through the lens of national security, the keynote explains how AI is now viewed as a new global arms race and why data centers, dependent on massive amounts of energy and critical materials, have become the limiting factor in AI advancement. This reality has profound implications for federal grant funding. As national priorities shift toward energy and critical materials, traditional funding pathways are giving way to agencies like the Department of Energy and the Department of War. The keynote challenges pre-award professionals, researchers, and institutional leaders to move beyond surface-level conversations about AI efficiency and instead recognize the deeper realignment underway. This watershed moment demands adaptation, reframed narratives, and a deeper understanding of how and why federal dollars will flow in the decades ahead.

9:15 - 10:45am Breakout Sessions

Location: 103AB, Floor 1
Presenters: Dr. Mark Van Dyke, University of Arizona

The proposal life cycle begins with project ideation and ends with submission to the sponsor. Along this well-traveled journey, there are ample opportunities to use AI-based tools to create efficiencies, but currently no commercial AI tools exist that purport to directly impact success rate (“hit/win rate” or the number of awards divided by total submissions). While a well-written proposal is expected to receive higher scores, proposers and sponsors alike are approaching the use of generative AI cautiously with both in agreement that AI should not be the primary author of a proposal. That said, AI is rapidly approaching human-level intelligence in complex tasks, including within the proposal life cycle, and will only get better. The College of Engineering at the University of Arizona recently began an effort to not only explore the use of AI in various steps of the life cycle but investigate whether a custom large language model (LLM) approach could provide a human-equivalent pre-review of human-authored proposals and thereby increase hit rate. The effort began with a college-wide questionnaire and focus group discussions to understand the current state of AI usage among faculty PI’s, as well as comfort level and attitudes toward expanded use. In the development phase, commercial LLM’s were tested for their abilities to perform several administrative tasks, as well as proposal pre-review. Evaluations of open-source (free) and subscription-based LLM’s were performed. Three different architectural approaches were also investigated: 1) as-is use of the commercial LLM, 2) fine-tuning the LLM with institutional and public data, 3) using a retrieval-augmented generation (RAG) layer of institutional and public data in combination with the commercial LLM, and 4) using ChatGPT’s built-in feature to create a custom GPT based on limited institutional and public data. Outcomes from two proposal submissions (pre-award phase only) will be discussed, one to the National Science Foundation and the second to the Department of Defense/War.<

Session Prerequisites:
Understanding of LLM customization. Knowledge of pre-award internal and peer review processes. Awareness of open-source and subscription-based LLM resources.

Location: 208B, Floor 2
Presenter: Matt G. Smith, Boise State University and Gigi Smith, College of Western Idaho

The rise of AI in research administration often sparks two conflicting reactions: excitement for efficiency and a defensiveness against AI’s impact on humans–on the work we do, the values that drive why we work, and how to articulate task automation alongside skills augmentation. In this two-part interactive session, Gigi and Matt Smith (Boise State University) move beyond high-level theory to unpack both the human side of change management and the practical side of AI tool creation.
Part I: Leading the Human Side of AI Implementation
We’ll begin with candid, field-tested lessons from implementing AI in a higher ed ecosystem, turning these experiences into collective takeaways for participants. By merging scholarship anchored in social sciences with research on barriers to AI advancement such as the common misfiring of distrust and uncertainty, the stage is set for evolving the skillsets leaders now need. Expect honesty, a bit of humor, and action plans for leading your team through technological change that invites a community-driven return to shared values to reset why we do the work.
Part II: Hands-On Workshop - Building an AI Assistant for Research Administration
Then we’ll roll up our sleeves. In this guided, live workshop, participants will use their laptop and preferred AI environment (e.g., Amazon Quick Suite, ChatGPT, Claude, Gemini, Copilot, an in-house platform) to build a working instance of the 324-CR Solicitation Checklist assistant. This is not a demo; it’s an actual build following the principle of human-in-the-loop design.

Learning Objectives:
Develop a barrier-breaking mindset to scale. You’ll acquire a strategy toolkit that overturns prevalent barriers in AI engagement to create capacity in the workplace for human-centric autonomy.
Learn how to build a blueprint for developing AI tools. You’ll create a custom AI assistant in your own environment using logic and design principles that function as a replicable blueprint for developing high-access, low-barrier AI tools.

Session Prerequisites:
Tech Requirements: Please bring a laptop with access to a paid/advanced AI tool with an assistants feature (e.g., Amazon Quick Suite “Agents,” ChatGPT “custom GPTs,”Claude “Projects,” Gemini “Gems,” Copilot “Agents,” an in-house platform equivalent) to participate in the workshop. (Note: If you do not have an AI assistant tool, we may be able to provide temporary access. Please email Matt Smith at mattsmith2@boisestate.edu as soon as possible if you have this need.)
Data Usage: This is an educational workshop. At the conclusion of the session, participants will be asked to remove specific, copyrighted workshop materials from their personal/institutional AI environments.

Location: 209B, Floor 2
Presenter: Ayham Boucher, Cornell University, Zachary Jacques, Cornell University and Keelan Schule, OpenAI

Most research administration teams have started experimenting with AI, but what does it take to move from experimentation to real adoption and impact? This session brings together AI experts from Cornell and OpenAI to share practical lessons from the front lines. The presenters will share how to approach upskilling staff at every level using tools like Claude Code to get real work done, when to build custom workflow applications, and how to make smart strategic choices with limited budgets. Attendees will see real examples of tools built in-house, including AI-driven foreign influence risk scoring, and hear directly from an OpenAI expert on tools, partnerships, and what’s coming next. The session closes with a look at how research administration roles across the award lifecycle are already beginning to transform, and what that means for the workforce ahead. The session also touches on the environmental sustainability implications of scaling AI across institutions

Learning Objectives:
1. Evaluate when to focus on upskilling staff using AI tools like Claude Code to get real work done versus investing in custom-built workflow applications, and understand the tradeoffs of each approach
2. Understand the infrastructure, partnerships, and prioritization decisions required to move from AI experimentation to real institutional impact on a limited budget
3. Envision how AI is beginning to transform roles and responsibilities across the award lifecycle, and what successful adoption means for the future of the research administration workforce
4. Learn from real-world examples and an OpenAI expert how to build and deploy AI solutions within research administration.

11:00am - 12:30pm Breakout Sessions

Location: 103AB, Floor 1
Presenters: Tomer du Sautoy, Atom Grants, Raphael Bernier, Atom Grants and Kym Fash, Denver Health and Hospital Authority

Artificial intelligence is rapidly becoming embedded in the daily work of academic researchers and administrators - from literature discovery and grant writing to collaboration mapping and proposal development. This session explores how AI is functioning as a "co-pilot" for research work, examining real-world applications, emerging archetypes of AI assistance (co-pilot, oracle, and actuator), and practical tools already in use across universities. We'll discuss semantic search technologies that break down disciplinary silos, demonstrate how AI can streamline grant discovery and writing while respecting funder guidelines, and honestly address where these tools fail. Attendees will leave with a clear understanding of what AI can do today, where it falls short, and how to implement these tools responsibly in their daily workflows.

Learning Objectives:
1. Understand the three emerging archetypes of AI in research: co-pilot (assistant), oracle (hypothesis generator), and actuator (autonomous experimenter), and identify which applications are ready for daily use.
2. Learn how semantic search technology works and how tools like Scite, Semantic Scholar, and Atom Grants can improve literature review, collaboration discovery, and grant matching.
3. Explore practical, compliant ways to use AI in grant writing and proposal development while adhering to NIH, NSF, and other funder guidelines.
4. Recognize common AI failures including hallucinations, bias, and the "knowledge problem" in scientific publishing, and develop strategies to mitigate these risks.
5. Assess the environmental and human costs of AI deployment and make informed decisions about responsible AI adoption in academic settings.<

Location: 208B, Floor 2
Presenters: Nathan Layman, University of Idaho and Katie Gomez Freeman, Southern Utah University

This hands-on session explores how to effectively deploy AI tools in research administration workflows. Participants will learn foundational prompt engineering techniques to get better results from AI systems, then progress to breaking down complex RA tasks into complete AI-assisted workflows. Through live demonstrations and practical examples, attendees will discover how to move beyond basic AI interactions to create efficient, repeatable processes for common research administration challenges.<

Learning Objectives:
Apply prompt engineering fundamentals to improve AI output quality and relevance
Break down complex research administration tasks into discrete, AI-manageable components
Develop complete AI-assisted workflows for common RA processes
Evaluate which tasks and workflows are suitable for AI deployment in their own institutions

Session Prerequisites:
Basic familiarity with AI tools like ChatGPT or Claude is helpful but not necessary.

Location: 209B, Floor 2
Presenter: Anthony O. Maceira, MZLS LLC, Mariola Abreu Acevedo, MZLS LLC, Marc A. Maceira, Honra and Vincent Borleske, University of Arizona

AI tools are becoming part of the sponsored research workflow — but the legal questions they raise are far from settled. Who owns content generated by AI in a grant proposal? What are the data privacy and confidentiality risks when AI processes federally funded research data? What do your sponsor agreements actually say, or fail to say. about AI use?
This session tackles these questions head-on from the perspective of practicing attorneys and a technology professional who work at the intersection of law, compliance, and innovation. The panel will walk through real legal risks institutions face today, explore how the evolving regulatory landscape, including new legislation and agency guidance, is shaping what's permissible, and offer practical takeaways for research administrators navigating this uncertain terrain.
Anthony Maceira is the Managing Member of MZLS and a certified CLE instructor on ethical rules of professional conduct. Mariola Abreu Acevedo is a Senior Associate at MZLS and former Assistant Solicitor General of Puerto Rico. Marc A. Maceira is the President of Honra and former Chief Solutions Architect at PRITS.
Expect a candid, practical discussion, bring your questions.

1:30 - 1:55pm Sponsor Demo Sessions

Location: 103AB, Floor 1
Presenter: Kathleen Halley-Octa, Attain Partners

As artificial intelligence and automation are poised to rapidly reshape research administration, institutions face growing pressure to modernize operations while safeguarding compliance and ensuring data integrity. Many universities, however, lack clarity on where—and how—AI can be effectively applied within research administration. This session introduces an AI Readiness Assessment framework purpose-built for research administration. The assessment evaluates institutional maturity across critical dimensions such as governance and policy, data readiness, research systems and infrastructure, workforce capabilities, and operational processes spanning the research lifecycle. By applying clear, maturity-based metrics, the framework enables institutions to move beyond isolated pilots and toward coordinated, sustainable AI adoption.

Learning Objectives:
1. Identify how AI readiness metrics can be used to assess institutional maturity and prioritize AI and automation investments within research administration.
2. Understand how a structured readiness approach reduces implementation risk while supporting compliance, data governance, and responsible AI use.
3. Apply readiness insights to improve pre- and post-award operations, enhance the researcher experience, and inform a long-term research administration modernization strategy.

Location: 208B, Floor 2
Presenters: Terry Durkin, Kuali

Identify common, real-world research administration challenges where AI can provide practical support, and see how those challenges are handled in Kuali.

Learning Objectives:
By the end of this session, participants will be able to:
1. Recognize common research administration challenges where AI can provide practical, day-to-day support.
2. Understand how AI is applied within Kuali Research and Kuali Build to assist with review, routing, and data analysis.
3. Evaluate where AI meaningfully reduces manual effort while maintaining compliance, transparency, and institutional oversight.
4. Ask informed questions about AI-enabled research systems to assess fit, risk, and readiness for their institution.

Location: 209B, Floor 2
Presenters: Aaron Wolpoff, Streamlyne

After a year marked by funding uncertainty and stretched research teams, many institutions are entering a rebuilding phase that demands smarter, not harder, approaches to obtaining research funding. Traditional methods like mass-distribution newsletters and bloated grant databases often generate more noise than clarity, using up time and energy while leaving the strongest opportunities buried. This session demonstrates how AI functions as a strategic partner in research development, and how to put it to use in a hands-on way. All attendees will catch a glimpse at the newest beta features of FundFit to truly understand how AI can help institutions pursue research funding with reassurance and confidence.<

2:05 - 2:30pm Sponsor Demo Sessions

Location: 209B, Floor 2
Presenter: Jason Seed, Endpoint IQ

More Information Coming Soon!

Location: 208B, Floor 2
Presenter: Sonia Singh, Huron and Christopher Crookston, Huron

Huron will demo their AI Awards+ tool which can support the award setup process in any system and facilitate date quality and cleanup. In addition, Huron will discuss how they are helping clients embed AI into core research managed services workflow, such as intake, triage, coverage analysis, budgeting, research finance, and contracting.<

2:45 - 4:15pm Breakout Sessions

Location: 103AB, Floor 1
Presenters: Jeff Warner, University of California Irvine

This session will discuss and demo the practical integration of Artificial Intelligence (AI) into your daily operations. Rather than just "adding another tool," we will explore how to use AI to be more responsive, data-driven, and efficient. Participants will see how to use AI to support communications, contract negotiations, and performance tracking.

Learning Objectives:
Participants will learn to evaluate and implement AI-powered communication strategies that reduce manual response times while maintaining professional standards in both internal and external stakeholder relations. Participants will be introduced to the potential for AI in contract negotiations, enabling efficient rationale development and leveraging institutional precedents. Participants will explore the use of generative AI to better understand and communicate metrics.

Session Prerequisites:
An interest in leveraging AI to improve workflows, reducing administrative burden by adopting new technologies that drive more efficient outputs, and developing tools to use for organizing and analyzing data.

Location: 208B, Floor 2
Presenters: Sonia Singh, Huron and Christopher Crookston, Huron and Dylan Ruediger, Ithaka S+R

As interest in AI adoption expands in research administration, many institutions struggle to move from experimentation and ideas to operational solutions. Successfully deploying AI in this context requires more than a subscription or a prototype. It depends on understanding how AI functions, how it interacts with institutional data, and how that capability aligns with research administration workflows. This session focuses on the practical realities of deploying AI in research administration, walking through key phases from early planning and process mapping to implementation, refinement, and operational use. Attendees will learn what to consider and what questions to ask before implementing AI, how to assess readiness across teams, and how AI capabilities can be translated into usable research administration workflows.

Learning Objectives:
Understand the key steps involved in moving AI from idea to implementation in research administration; Identify the questions and considerations that support effective AI deployment across teams and workflows.

Location: 209B, Floor 2
Presenter: Tanta Myles, Georgia Institute of Technology, with Industry Panelists including Andrea Christelle, University of Kentucky, Stephanie Gray, University of Florida and Dr. Mark Van Dyke, University of Arizona

As generative AI becomes a mainstay in the professional toolkit, funding agencies are rapidly evolving their disclosure requirements. This panel brings together leadership from multiple institutions to examine the challenges of AI transparency and disclosure. We will move beyond the simple question of "To use or not to use?" and instead focus on how, where, and why to disclose AI assistance to maintain integrity without sacrificing innovation. Specific topics will include current federal agency policy related to the use of AI in proposal development. Panel members will also share recommendations for communicating new federal requirements to facuty and staff, share interpretations of "substantive" use of AI providing insight on distinquishing between administrative editing versus content generation, and sharing any institutional guidelines or best practices to navigate conflicting sponsor guidelines.

Learning Objectives:
Build awareness of current federal policy and expectations of the use of AI in proposal development. Identify best practices for informing faculty of proposal requirements including ramifications of not disclosing the use of AI in the proposal submission process. Identify potential strategies for developing a busines process or department, division, or institutional policy to identify and monitor compliance with AI disclosure requirements.