- Study Finds Most Organizations Are Not Consistently Measuring AI Value Against Costs — A new KPMG study covered by Campus Technology finds that while AI investment keeps climbing, relatively few organizations consistently measure whether the technology's benefits actually justify its costs. The gap between spend and measurement is becoming a governance problem: budgets are being committed to AI faster than institutions can demonstrate returns. For higher education, where every dollar competes with teaching and research missions, unmeasured AI spend is especially hard to defend.
📌 Key takeaways:
- Campus IT leaders should insist on a standard value-per-cost measurement for every AI deployment before renewal conversations, not after.
- AI budgets that lack ROI evidence are the first ones cut when institutional finances tighten — building the measurement discipline now protects the program later.
- 🏛️ UCSD angle: UCSD's AI program routes usage through a central gateway, giving it the usage and cost telemetry most organizations in the study lack — the foundation for exactly this kind of value measurement.
- Nearly half of employers say AI will make college degrees more important — A new Gallup and Lumina Foundation poll finds 48% of employers believe AI will make college degrees more important, with nearly half expecting a degree to matter more in an AI-driven economy. Employers also expressed more confidence in higher education than the general public does, and most say they prefer hiring workers with a college degree even for roles that don't formally require one. It's a counterintuitive data point for a sector used to defending its relevance against "AI makes college obsolete" narratives.
📌 Key takeaways:
- Institutions should treat employer sentiment as a leading indicator for curriculum and credential design — the market is signaling durable demand for degrees that incorporate AI fluency.
- AI literacy embedded in general education is becoming an enrollment argument, not just a skills argument.
- Claude AI now available to Emory University community — Emory University announced campus-wide access to Claude AI for its students, faculty, and staff, making it one of the largest single-vendor university AI deployments this fall. The rollout is paired with Emory's published Responsible AI Guidelines and a structured license provisioning process, and it lands alongside Emory and Georgia Tech's combined $1B+ AI investment announced earlier this month. The move reflects a decisive shift from AI pilots to institution-wide, contractually governed deployments.
📌 Key takeaways:
- Campus-wide vendor deployments work best when paired with published responsible-use guidelines and a governed provisioning process — Emory is modeling both.
- Enterprise AI agreements are becoming strategic institutional investments on par with research computing, as the Georgia Tech–Emory billion-dollar commitment shows.
- 🏛️ UCSD angle: UCSD's AI platform takes a model-agnostic approach — routing multiple frontier models through a single campus gateway — rather than standardizing on one vendor, keeping leverage as peer institutions sign single-vendor deals.
- CWRU Partners with Google to Expand Campus-Wide AI Access with Gemini Enterprise — Case Western Reserve University announced a university-wide Google partnership bringing Gemini Enterprise accounts to all faculty, staff, and students via single sign-on, alongside no-code agent building and cross-system search across Google Workspace and Microsoft 365. The deal includes early researcher access to experimental tools like Co-Scientist, AlphaEvolve, and Project Aura, plus co-located Google engineers in a planned AI Discovery Hub. Notably, CWRU paired the announcement with explicit guidance steering users away from unapproved consumer AI accounts and restricting sensitive data entry.
📌 Key takeaways:
- The security posture matters as much as the product: CWRU's deployment explicitly forbids training on university data and redirects users from consumer accounts to the governed enterprise tier.
- Vendor partnerships that include embedded engineers and early access to experimental research tools change the calculus beyond simple license pricing.
- Campuses evaluating similar deals should watch the integration claims — unified search across Workspace and Microsoft 365 is the kind of promise that deserves a pilot before a promise to faculty.
- Advancing coding in low-resource programming languages — OpenAI published new work improving ChatGPT's performance in low-resource programming languages, including new evaluations for software engineering, SQLite, and COBOL migration. For organizations sitting on decades of mission-critical legacy code, AI-assisted migration of aging COBOL systems is moving from theory to a credible modernization path. Higher education is full of exactly these systems — financial records, payroll, and student information platforms still running on code older than most of the staff maintaining it.
📌 Key takeaways:
- Institutions with legacy administrative systems should begin inventorying COBOL and other aging codebases now — AI-assisted migration tooling is maturing faster than most modernization roadmaps assume.
- The shrinking pool of retirement-age programmers who know these languages makes AI-assisted migration a continuity issue, not just a cost issue.