In October 2026, Engr. Axel B. Banay, LMI's Founder and CEO, led an afternoon working session for the leadership and deans of a private university. It was called AI with Purpose, and it was built around one aim: leave the room with a clear decision, not just a discussion.
To protect the university's privacy, we do not name it here. Its self-ratings and the ideas its teams proposed stay with the university.
Why now
In September 2026, CHED issued Memorandum Order No. 21, series of 2026, a national framework for artificial intelligence in higher education. It does not ban AI. Instead, each school writes its own policy, students disclose when they use AI, and people, not machines, stay in charge of decisions such as grading, admissions and scholarships. Meanwhile, anyone on campus can already open a free AI tool today. The real question is whether that use is governed.
How the afternoon worked
1. A readiness check across six pillars
Leaders rated their institution from 1 (unaware) to 5 (leading) on six pillars that decide whether AI succeeds: mission and strategy, governance and policy, data and privacy, infrastructure and security, people and literacy, and use cases and value.
2. Four tables, one question
Deans led four tables (teaching and learning, health sciences, student services and administration, and business, hospitality and basic education) around three questions:
- Which task eats the most hours every week?
- Where do students wait, get stuck or drop out?
- What would you never let AI decide?
3. Sort by impact and effort
Every idea went on the wall and was sorted by impact and effort. Quick wins come first. Ideas that affect a person's academic outcome, such as AI in admissions, grading or early-warning systems, wait until governance and human review are in place.
Five house rules any school can adopt this week
- No personal data in public AI tools, whether student, patient or employee.
- No confidential files in public AI tools, such as exams, grades, contracts or minutes.
- A person checks and owns every AI output that gets used.
- Disclose AI help in official documents and in student work.
- AI detectors are a clue, never proof of misconduct.
Foundation first, then pilots, then scale
- Months 0 to 2: assess. An eight-week readiness assessment and a gap check against CMO No. 21.
- Months 2 to 6: foundation. A small governance committee, an approved AI policy and tools list, AI literacy training for faculty and staff, and stronger account security.
- Months 6 to 12: pilot. Three measured pilots, starting with low- or medium-risk uses.
- Years 2 to 3: scale. Grow what worked, and keep sensitive data on platforms the school controls.
Privacy by design
AI should never be the reason personal data leaks. The session grounded every step in the Data Privacy Act of 2012 (Republic Act No. 10173) and the National Privacy Commission's Advisory No. 2024-04 on AI: a lawful reason and the least data needed for every AI use, openness about where AI is used, a person in control of decisions, and logs good enough to report a breach within 72 hours.
What comes next
The afternoon closed with four decisions for the university's leaders: name an executive sponsor, form the committee, adopt the house rules, and decide on an eight-week readiness assessment. That assessment ends with a readiness scorecard, a CMO No. 21 gap analysis, a draft AI policy, the top ten use cases with their value and risk, a three-year roadmap and an AI risk register.
"AI should strengthen people, never replace them."
Working out where AI fits in your school or business? Book a conversation or email lori@onelmi.com.