Power Isn't Progress: What Should AI Be Teaching Higher Education?
Featuring Blessed Pepple, Technical Advisor, Helix4HE Advisory Board
Artificial intelligence has rapidly expanded what students, educators and institutions can produce. It can generate text, analyse data, summarise research, translate content, write code, personalise information and automate processes that once required considerable human effort.
The capabilities are extraordinary. Yet higher education needs to ask a more fundamental question:
What should working with AI actually be teaching us?
The distinction matters because efficiency and progress are not the same thing. A university can become more efficient without becoming more educationally effective. A student can produce more without becoming more capable. An institution can introduce sophisticated technology across teaching, recruitment and administration without fundamentally transforming the experience of learning.
The risk is that increased output begins to look like increased capability.
This is a distinction that Blessed Pepple, Technical Advisor on the Helix4HE Advisory Board, has encountered repeatedly throughout a career at the intersection of higher education, technology and institutional transformation.
Blessed is an international EdTech strategist and specialist in AI-driven higher education transformation. He works with ApplyBoard as an Associate Account Director, a role that places him close to the rapidly evolving relationship between technology, international student recruitment, institutional partnerships and the international student journey. His wider experience spans digital transformation, student engagement and the use of technology to support institutional performance and access to education.
Reflecting on successive waves of technological change, Blessed observes:
“Across my career, I've seen the same pattern play out with new technology. People get excited about what it can produce, and forget to ask what it should be teaching them.”
That is a useful starting point for higher education. Much of the current conversation about AI has understandably centred on capability: What can it do? What can it automate? What can it generate? How much time can it save?
Those are important questions, but they are not sufficient.
As Blessed puts it:
“AI is a good example. It's incredibly powerful, but power isn't progress.”
Power tells us what technology is capable of doing. Progress requires us to decide whether its use is actually improving learning, strengthening judgement, enhancing institutional performance or creating better outcomes for students.
And that takes the debate beyond technology itself.
When Knowledge Becomes Easier to Access, Human Capability Matters More
Universities have never existed simply to provide access to information. Nevertheless, access to specialist knowledge, academic expertise and scholarly resources has historically formed an important part of the value of higher education.
That environment has changed profoundly.
A student can now ask an AI system to explain a theory, compare competing arguments, summarise research, translate complex material, structure an essay, generate code or produce an apparently coherent response within seconds.
The existence of those capabilities does not make higher education less important. It changes where much of its value increasingly lies.
Blessed identifies the challenge clearly:
“Higher education is a critical sector for this question, since we're preparing people to join the workforce. AI is an incredible opportunity for innovation, but it carries real risk. When knowledge is this accessible, you have to ask what a degree is actually worth.”
The answer cannot simply be the transmission of more information.
As information becomes easier to access, the ability to interpret, interrogate and apply it becomes more important. Students need to judge whether an answer is credible, recognise weak evidence, distinguish confidence from accuracy, identify bias, detect what is missing and make decisions when information is incomplete or contradictory.
They need to explain those decisions, defend them and remain accountable for them.
This is why Blessed’s emphasis on human capability is so significant:
“What actually moves organisations, institutions and students forward is human capabilities like judgement, accountability, self efficacy, curiosity, adaptability.”
These should not be regarded as peripheral employability skills sitting somewhere around the edge of the curriculum. In an AI-enabled world, they move much closer to the centre of what an educated graduate needs to demonstrate.
There is an important paradox here. As technology becomes more capable, some distinctly human capabilities may become more valuable, not less.
AI can produce persuasive answers without understanding their consequences. It can generate options without bearing responsibility for which one is chosen. It can accelerate analysis without possessing professional judgement.
Humans still carry those responsibilities.
For graduates entering healthcare, engineering, finance, education, law, management, public administration or any other profession where decisions have consequences, accountability cannot simply be transferred to the system that helped generate an answer.
AI literacy must therefore mean considerably more than knowing how to operate a platform. It must include knowing when to question an output, how to verify it, how to contextualise it and when human judgement needs to override machine suggestion.
The Real Question Is Whether Students Are Learning With AI
This brings us to perhaps the most important distinction in Blessed’s contribution:
“The real question for universities is whether students are learning with AI, or just churning out work with it.”
That question goes directly to the future of assessment.
A polished piece of work no longer necessarily tells us how much learning occurred in producing it. An essay may represent extensive reading, synthesis, questioning and independent judgement. It may equally represent an effective prompt followed by relatively limited human intervention.
The finished product can look remarkably similar.
That should encourage institutions to reconsider what assessment is actually intended to reveal.
If an assessment primarily measures whether a student can produce an output that AI can now generate convincingly, the problem may not simply be inappropriate use of technology. The assessment itself may need to evolve.
Blessed makes the connection explicit:
“That comes down to how we assess teaching and learning, and whether we're building the skills and human capability students need to be transformational with this technology, rather than treating it as an assistant that thinks for them.”
The long-term answer is unlikely to be an endless contest in which universities design tasks that AI cannot complete and technology companies subsequently develop systems capable of completing them.
Nor does it make sense to prepare students for contemporary professional life by pretending these technologies do not exist.
A more productive question is how assessment can make thinking visible.
Why did the student accept one source and reject another? What assumptions did they identify? What did they verify independently? Where did AI contribute? Where was it wrong? What did they challenge? How did their thinking change? What judgement did they ultimately exercise?
This could increase the importance of oral defence, authentic projects, simulations, live problem-solving, iterative assessment, reflective analysis and tasks in which students critique or improve AI-generated material.
It may also require institutions to become more explicit about AI use rather than simply attempting to drive it underground.
A student who can explain where AI was used, what it suggested, what they rejected and why may reveal considerably more learning than someone who simply submits a polished final product.
The educational value shifts from the answer alone to the reasoning behind it.
This is where the difference between technological efficiency and educational transformation becomes especially important.
“Used properly, AI should push us towards transformative change, not just efficient output. That's the distinction I hold onto in everything I build.”
Introducing AI does not, in itself, transform a university.
An administrative process may take ten minutes instead of an hour. A lecturer may create materials more quickly. A marketing team may produce more content. Students may receive responses almost instantaneously.
All of those developments can be valuable.
But transformation requires deeper questions.
If students can obtain explanations of foundational knowledge on demand, how should teaching time be used? If AI can produce a competent standard essay, what should academic writing assessment now measure? If information can be personalised instantly, how should curriculum design respond? If routine administrative processes can be automated, where might universities deliberately preserve or increase human interaction?
And if machines are capable of undertaking more routine cognitive activity, what higher-order capabilities should graduates possess?
These are questions of educational design and institutional purpose, not simply technological adoption.
From Recruitment Efficiency to Better Student Fit
The same principle extends beyond teaching and assessment.
Blessed’s experience in international education gives particular relevance to another rapidly developing area: the use of AI in student recruitment.
AI is increasingly influencing how prospective students discover institutions, compare opportunities, interact with recruitment and admissions services and make decisions about where to study.
The immediate institutional opportunity is obvious. AI can personalise communications, automate follow-up, respond to enquiries rapidly, support admissions processes and help recruitment teams manage larger volumes of prospective students.
But efficiency should not necessarily be the final objective.
Blessed argues:
“The same applies to international recruitment. Used well, AI can help prospective students find the right opportunity for them, not just optimise conversion and admissions, but genuinely place students where they belong.”
That shifts the conversation from recruitment efficiency towards educational fit.
For an international student, choosing a university can involve considerable financial, emotional and personal commitment. Students may be comparing countries, institutions, programmes, living costs, visa environments, employment opportunities, languages and cultures while making decisions from thousands of miles away.
Poor fit can carry significant consequences for both the student and the institution.
AI could therefore be used not simply to persuade a prospective student to apply, but to help them understand whether a programme genuinely aligns with their ambitions, academic preparation, financial circumstances, preferred learning environment and longer-term career goals.
It could help students ask better questions and compare opportunities more intelligently. It could support institutions in identifying when a different programme may be more appropriate. And it could help create more realistic expectations before enrolment.
That represents a very different conception of intelligent recruitment.
The objective shifts from transaction towards guidance, with the potential to support stronger retention, progression, wellbeing, satisfaction and graduate outcomes.
The same distinction can be applied across the institution.
In recruitment, AI can process more enquiries, or it can help students make better choices.
In teaching, it can produce more content, or it can create space for deeper interaction.
In assessment, it can generate answers, or it can become something students are expected to interrogate critically.
In administration, it can reduce workload, or it can release staff to spend more time on the complex human interactions that cannot easily be automated.
The technology does not determine which version emerges.
Institutional choices do.
Measuring What Actually Matters
This leads to a wider question: how should a university determine whether its adoption of AI is actually succeeding?
Numbers of licences, active users, automated processes, hours saved, pieces of content produced or applications converted can tell an institution something about adoption and efficiency.
They reveal much less about transformation.
More meaningful questions might be whether students are demonstrating stronger judgement, whether assessment is producing deeper learning, whether staff are retaining professional accountability while using AI, whether students are becoming more independent rather than more dependent, and whether prospective students are making better-informed choices.
Institutions should also ask whether technology is improving access without removing human interaction where that interaction matters most.
And ultimately, whether graduates are genuinely better prepared for workplaces in which human and artificial intelligence will increasingly operate together.
AI can easily be framed either as a threat to higher education or as a solution to its problems.
Neither position is sufficient.
AI is a powerful technology whose educational value will ultimately depend on the choices institutions make around it.
That creates an opportunity to revisit questions that were arguably overdue for reconsideration anyway: what assessment is for, how learning should be demonstrated, what graduates should be capable of, where professional judgement must remain human, and how recruitment might move beyond conversion towards better student fit.
The institutions that respond most effectively may therefore not be those that introduce AI fastest.
They may be those that understand most clearly what they want people to become capable of because AI exists.
Blessed’s central distinction is worth returning to:
“AI is a good example. It's incredibly powerful, but power isn't progress.”
The measure of successful AI adoption in higher education should not simply be whether students, staff and institutions can produce more.
It should be whether people can think better, judge better, adapt better and take greater responsibility for the decisions they make.
The opportunity for higher education is not simply to use increasingly powerful technology.
It is to use that technology to develop increasingly capable people.
Disclaimer: This article presents an informed perspective on AI and education and should not be treated as a substitute for institution-specific professional advice.
Comments