UP Law Restricts AI in First-Year Classes, Raising a Bigger Question About How Students Learn to Think

MANILA — As artificial intelligence becomes increasingly embedded in education, the University of the Philippines College of Law is taking a deliberately different approach in some of its most foundational classrooms.

Beginning Academic Year 2026–2027, the UP College of Law is restricting the use of artificial intelligence tools and digital or electronic devices in onsite first-year Juris Doctor core classes.

The move is notable not simply because one of the Philippines' leading law schools is placing limits on AI. It is notable because of the reasoning behind it.

Faculty members have raised concerns over the quality of students' work, particularly in areas such as writing, comprehension and independent analysis. For first-year law students, these are not peripheral academic abilities, but foundations upon which legal education is built.

And that makes UP Law's decision part of a much larger debate now emerging across higher education.

The question might not be about whether universities should adopt artificial intelligence anymore, but rather when students should be required to think independently of it.

A Boundary Is Being Drawn.

It would be easy to interpret restrictions on AI as resistance to technological change. But that misses an important distinction.

Universities are increasingly recognising that AI literacy and independent intellectual capability are not necessarily competing objectives.

Students will undoubtedly enter workplaces where artificial intelligence is commonplace. Lawyers will use AI-assisted research. Businesses will automate analysis. Researchers will work alongside increasingly sophisticated computational tools.

Knowing how to use these systems responsibly will therefore become an important professional skill.

However, universities face another responsibility which is to ensure that students can perform the underlying intellectual work themselves.

A calculator is useful because the person using it is still expected to understand mathematics.

Translation software is more valuable when its user understands language. AI-assisted legal research becomes more reliable when the lawyer using it can recognise weak reasoning, missing context or an incorrect interpretation of the law.

The same principle may increasingly apply to generative AI.

Before students learn to accelerate their thinking, universities may first want evidence that the thinking exists.

The Philippine Context Makes the Decision More Interesting

UP Law's position comes at a time when the Philippines is moving broadly toward greater AI adoption.

In February, the Department of Education issued foundational guidelines permitting AI use in public schools provided that it remains ethical, pedagogically appropriate and human-centred. Under the policy, AI is intended to function as an auxiliary tool rather than replace teachers' judgment.

The government has also been expanding AI literacy initiatives, teacher training and AI-supported education systems as part of a broader effort to prepare Filipino learners for an increasingly technology-driven economy.

There is no inherent contradiction between these approaches.

In fact, together they may illustrate a more mature stage of AI adoption.

A national education system can promote students' understanding and use of artificial intelligence while individual institutions choose to keep specific environments such as examinations, foundational courses, writing exercises, oral assessments, or first-year programmes intentionally less automated.

The emerging policy question is therefore becoming more nuanced than simply being "for" or "against" AI.

It is about deciding where AI adds educational value and where removing it may create educational value.

First Year Matters

The focus on first-year law students is particularly significant.

Legal education depends heavily on reading difficult material, identifying relevant facts, interpreting judgments, constructing arguments and defending conclusions.

These abilities are developed through repetition.

A student struggling through a complicated judgment may take considerably longer than an AI system to summarise it. But the struggle itself can be part of the education.

The same applies to writing.

Constructing an imperfect argument, discovering weaknesses in it and rewriting it may appear inefficient compared with asking an AI system to restructure the reasoning.

Educationally, however, that inefficiency can be productive.

Universities have historically been places where students are required to practise difficult intellectual tasks precisely because those tasks eventually become easier through mastery.

AI introduces an unusual challenge because it can remove some of that difficulty before mastery has developed.

AI Detection to Assessment Redesign

For the past several years, much of the university conversation around generative AI has focused on academic integrity including plagiarism, disclosure rules, AI detection and the permissibility of specific tools.

The next phase may be considerably more interesting.

Universities may begin redesigning assessments around what they actually want students to demonstrate.

Some tasks may openly allow AI and evaluate how intelligently students use it. Others may deliberately remove digital assistance and test writing, reasoning, recall, interpretation or oral defence independently.

The distinction could eventually become an important marker of academic quality.

Instead of assuring students that they will have unlimited access to technology, universities might need to clarify the capabilities their graduates can show without technological assistance and the capabilities that are improved by using it.

That distinction could influence curriculum design, assessment methods and potentially even admissions.

AI Discussions May Include Academic Quality

There is also a reputational dimension.

Universities have spent years promoting digital transformation, smart campuses and technology-enabled learning as evidence of modernity.

If employers start to question whether graduates can write, analyze information or solve problems independently, institutions may need to show something different by demonstrating that technology has improved learning while still fostering the development of fundamental skills.

That could eventually affect how universities communicate academic quality.

The strongest institutional position may therefore be neither "AI everywhere" nor "AI nowhere."

It may involve the ability to explain precisely where AI fits into the learning process and where students are still expected to engage in the intellectual heavy lifting themselves.

The Debate on Student Use of AI Advances

UP Law's restrictions are unlikely to settle the wider debate over artificial intelligence in universities.

They do, however, illustrate how that debate is changing.

The initial stage of generative AI in education was characterized by a reactive approach. Institutions were concerned about cheating, students explored new tools, and faculty formulated policies as the technology advanced more rapidly than academic regulations could adapt.

The next phase may require something more deliberate.

Universities will have to decide which human capabilities remain essential even when technology can perform similar tasks faster.

For law, that may mean reading carefully, constructing arguments and defending them independently.

For other disciplines, the answer will be different.

However, the underlying question is likely to gain importance in higher education.

Before students learn to work with artificial intelligence, what should they first demonstrate they can accomplish without it?

That may ultimately be a more important measure of AI readiness than how quickly universities adopt the technology.