CQUniversity researchers warn of rise of ‘synthetic competence’ in AI age
CQUniversity researchers are warning that generative artificial intelligence may be creating a new class of ‘synthetic experts’ – people who can appear highly capable through AI-generated work without developing deep knowledge, judgement or genuine expertise.
The concept is outlined in a newly published article by CQUniversity researchers Associate Professor Meena Jha, Professor Stephen Colbran and Professor Ken Purnell, which argues that higher education must remain focused on developing and certifying genuine human expertise rather than AI-assisted performance.
Professor Purnell said while much of the public discussion around AI had focused on plagiarism, assessment and productivity, a more important question was emerging.
“Most debate focuses on tools, plagiarism and assessment, yet largely ignores how AI is reshaping the development of expertise itself,” Professor Purnell said.
“We wanted to shift the conversation from ‘How do we use AI?’ to ‘What does it mean to become an expert in an AI-saturated world?’”
In the paper, the researchers describe synthetic competence as the appearance of high ability created by AI systems without the underlying human understanding or skill traditionally associated with expertise.
Professor Purnell said increasingly sophisticated AI tools had made it easier than ever for people to produce polished work that appeared authoritative.
“With GenAI, it is now easy to sound knowledgeable while remaining a novice,” he said.
“The risk is that employers, professions and the public may trust people who can produce impressive documents but cannot reason through complex problems, assess evidence, or take responsibility when things go wrong.”
Central to the paper is the concept of biological memory, described as the network of knowledge, skills and judgement that are developed through learning, practice, feedback and reflection over time. The article argues that genuine expertise relies on these internal cognitive structures, rather than access to information or being able to prompt AI-generated answers.
“Expertise is not just recall; it is the capacity to recognise patterns, make sound judgements, and improvise under pressure,” Professor Purnell said.
The researchers argue that prompting AI systems should not be confused with true learning and warn that excessive reliance on AI could reduce opportunities for the types of thinking that help build lasting knowledge and professional judgement.
“Prompting is asking a machine to think for you; learning is changing your brain so you can think for yourself,” Professor Purnell said.
“If universities blur that distinction, we risk mistaking strong prompting skills for deep understanding and awarding degrees for outsourced cognition rather than genuine expertise.”
Rather than arguing against the use of AI, Professor Purnell said universities should focus on helping students use the technology in ways that strengthen learning. Suggested approaches include asking students to critique and verify AI-generated content, explain their reasoning, and demonstrate how their understanding has developed over time.
According to Professor Purnell, the challenge for higher education in the new AI age is to ensure graduates can work effectively with AI.
“The core message is simple: universities must remain in the business of cultivating and certifying genuine human expertise, not merely certifying AI-assisted performance,” he said.
“Successful adoption would mean universities will see graduates who can collaborate with AI while maintaining strong independent judgement, deep disciplinary understanding and ethical responsibility.”
The article, Beyond Prompting: Biological Memory, Cognitive Offloading, and Human Expertise in the Age of GenAI, was published in Frontiers in Educational Practice and Research on 2 July 2026.
