By Jason Oliver, Partner, Madeleine Scott, Lawyer and Jarod Aitchison, Paralegal

The recent decision of Lenehan J in Chief Executive Officer of the Tertiary Education Quality and Standards Agency v Chegg, Inc. [2026] FCA 330 is the first to consider in any depth the meaning of the defined expression “academic cheating service” in the Tertiary Education Quality and Standards Agency Act 2011 (Cth) (TEQSA Act). The case is notable for what it reveals about the limits of the TEQSA Act in the age of generative artificial intelligence.

Section 114A of the TEQSA Act, introduced by amendment in 2020, is a criminal and civil penalty provision which prohibits persons from providing, or offering to provide, academic cheating services.  An “academic cheating service” is the provision of “work” to, or the undertaking of “work” for, students, in circumstances where the work is (or could reasonably be regarded as being) or forms (or could reasonably be regarded as forming) a substantial part of an assessment task that a student is required to undertake personally.

One issue in Chegg was the meaning of “work” for the purposes of that definition.  The TEQSA Act does not define “work”, so the Court had to discern its meaning in accordance with well-established principles of statutory construction, beginning and ending with the statutory text, understood in context and in light of its statutory purpose.

At paragraphs 51 to 54 and 71 of the decision, Lenehan J accepted the submissions of both parties that “work” in s 114A means the intellectual labour or effort of the relevant person, or at least the product of such labour or effort.  The material supplied by Chegg (being prepared by human subject-matter experts in response to student questions) fell squarely within the scope of the provision.

That conclusion is perhaps unsurprising given the technological landscape that existed at the time s 114A was enacted, but it reveals an apparent lacuna in the age of generative AI.  As the Court acknowledged, s 114A was enacted because “Academic cheating services seriously undermine the integrity of the Australian higher education system and the consequences of students accessing these services can be severe.  If a student graduates having relied on the assistance of a cheating service rather than personally achieving the required level of academic attainment, in fields such as medicine or engineering, the consequences can be life threatening…”.  That mischief is no less acute in an age where students can obtain bespoke answers instantaneously through generative AI tools.  But if “work” is confined to human intellectual labour or its product, then material generated entirely by automated means may fall outside the scope of s 114A.

The practical significance of this issue is underlined by the Court’s observation at paragraph 95 that Chegg’s business has suffered a sharp decline in recent years.  That decline was one of the factors taken into account by the Court in deciding what civil penalty to impose.  The decision does not identify the cause of the decline, but Chegg has attributed it to the rise of generative AI.

The above begs the question: how, if at all, should the legislation be updated to address the risk of generative AI supplanting the human misconduct which s 114A was enacted to outlaw?

The Australian Government and TEQSA, in conjunction with the universities, are continuing to develop their responses to the risks to academic integrity posed by generative AI.  To date, the focus has been on addressing pedagogical, research and assessment design risks, including through the development of responsible use guidance and enforcement by universities of academic policies and procedures at the student level.

It remains to be seen whether the Australian Government will attempt to amend the TEQSA Act to somehow bring AI-driven academic cheating services within the scope of s 114A.  On the one hand, left in its current form, there is a risk that s 114A will quickly become obsolete as a means of tackling the mischief described above, or else only operate to expose the few remaining human service providers to severe punishment while AI-driven service providers may avoid liability.

On the other hand, there may be no straightforward way to bring it “up to date”.  For example, any reform is unlikely to be as simple as amending the definition of academic cheating service to remove the requirement of human labour or effort.  Such an approach, without more, would theoretically expose all generative AI providers to a civil penalty, since they are offering to provide a generative AI product which is capable of being used for academic cheating.