AI in RA
« AI in RA

#20 — Ethics boards deploy AI, OpenAI Plus grants, Georgia Tech Phase 2

November 18, 2026

Sources

  1. AI Could Speed Up Research Ethics Approvals in 2026
    * AI systems are now assisting research ethics committees in reviewing study proposals, reducing approval times significantly. * The technology analyzes protocols for compliance issues, flags potential ethical concerns, and streamlines administrative workflows. * Early adopters report approval times dropping from several weeks to just days while maintaining review quality.
  2. OpenAI expands free AI access to 200,000 researchers
    OpenAI will open 50,000 new research grants on November 23, offering Plus access for recipients and up to two collaborators each.
  3. Georgia Tech Researchers Develop AI Tool to Make Internal Information Easier for Companies to Find
    Georgia Tech’s Tech AI initiative and Velastegui Ventures are developing an enterprise AI knowledge platform designed to help organizations search and interact with internal institutional information using natural language queries. The platform combines retrieval-augmented generation, semantic search, and citation-backed responses to retrieve and contextualize data from documents, websites, and databases. Following the completion of Phase 1's technical foundation, the project is moving into Phase 2 to emphasize scalability, governance, monitoring, and broader enterprise adoption.
  4. Is It Safe to Upload Grants to AI? Rules by Account
    How ChatGPT, Claude, Gemini and Copilot handle uploaded files by account type, which funder rules limit what you may upload, and a safer workflow for grant documents.

Also this week

Full transcript
Can algorithmic systems evaluate research protocols and identify compliance issues before human committees step in? That is what we are tracking today on AI in RA, your regular update on artificial intelligence in research administration. Here is the report. Research ethics committees have begun deploying artificial intelligence tools to screen study submissions before formal reviews take place. The reported result is a drop in turnaround times from multiple weeks to several days, targeted primarily at administrative bottlenecks rather than altering the evaluation criteria. The software runs pre-screening for regulatory compliance and flags procedural issues before documents ever reach committee members. At the same time, feeding institutional documents into external systems is drawing scrutiny. Platforms like ChatGPT, Claude, Gemini, and Copilot maintain different data retention and processing policies depending on the subscription tier. And grant funding agencies enforce strict bans on uploading confidential text or proprietary research proposals into commercial systems. Which requires institutions to construct data sanitization protocols and internal screening steps prior to any external upload. Access points to these models are expanding anyway. OpenAI announced 50,000 research grants opening for application on November 23. Each grant provides ChatGPT Plus access and allows the recipient to provision seats for up to two collaborators for text generation, data synthesis, and workflow management. Alongside individual researcher access, centralized campus deployments are progressing. Georgia Tech’s Tech AI initiative and Velastegui Ventures transitioned their enterprise knowledge platform into Phase 2. They completed the core technical build in Phase 1, using retrieval-augmented generation, semantic search, and automated citations across institutional repositories and databases. Phase 2 shifts the focus toward operational governance, system monitoring, scaling, and cross-department administrative rollout. That focus matches broader software management trends. A comparative evaluation recently examined ten artificial intelligence agent frameworks, looking closely at operational controls, monitoring capabilities, lifecycle requirements, and cost structures. We will be back next week with more developments in research policy, tools, and governance. From AI in RA, thanks for listening.

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 →