Authors

Professor Sandie Dann,
Professor Michael Draper
and Dr Irene Glendinning,
QAA Academic Integrity Advisory Group members

Earlier this month, the European Network for Academic Integrity (ENAI) held its 9th Annual Conference on Ethics and Integrity in Academia at the University of Derby, the first time the conference had been held in the UK. ENAI is a membership association and the largest academic integrity network in Europe, engaging with the academic integrity community worldwide. Over 200 delegates, staff and students attended from across Europe and further afield including Canada, Australia, India and Nepal, demonstrating the international reach of the conference and the partnerships that have been developed by ENAI.

This blog, contributed by three members of the QAA Academic Integrity Advisory Group, shares the conference highlights and takeaways.

During the event, another member of the Advisory Group, Dr Thomas Lancaster, was presented with the prestigious Tracey Bretag Memorial Award recognising lifetime achievement in academic integrity. QAA warmly congratulates Thomas on this award which recognises his pioneering work on contract cheating and tireless campaign to highlight opportunities and risks to integrity posed by artificial intelligence.

Detection tools, artificial intelligence and text spinners

Significant debate was generated by the session on artificial intelligence (AI) detection tools based on recent research by a team of ENAI members concluding that they don’t work reliably. Context, of course, is everything. The research methodology and associated results were explained with a clear message that detection tools results should not be used as the sole evidence in academic misconduct procedures as they are not accurate or reliable. Using text spinners (tools that adjust wording in an effort to evade being detected as plagiarism) on AI generated material was demonstrated to be a sure-fire way of fooling AI detection tools. Weaknesses in this area are being further investigated by commercial vendors and academics who are developing new tools rapidly.

Rethinking assessment design

Revising assessments to deter AI misuse was highlighted by many presenters, although one delegate noted concerns that revising assessments will have an impact on staff workload and so requires careful thought.

Special attention was drawn to the need for addressing marking and moderation burdens, particularly having regard to the worsening position of staff-student ratios at some institutions and the use of fractional staff and research students in marking duties. Knowing your students has previously been identified as a way to prevent impersonation and contract cheating, but the drive towards anonymous marking makes it difficult to check that work submitted for assessment is authentic.

It was however noted that concerns about ensuring secure assessment were driving higher education providers to return to the examination hall. Many participants reflected that the innovations created during COVID were being rolled back in view of concerns about academic rigour and maintaining educational standards. This highlights the need for more training and support for staff to produce authentic assessment which requires students to demonstrate critical analysis and interpretation rather than recall. However, there can be challenges with this approach arising from the need to meet some professional, statutory and regulatory body (PSRB) requirements which may still focus on what students 'know', along with a desire to retain closed book, timed exams.

Defining generative artificial intelligence

There was a view amongst delegates that institutions were not agile in their response to AI. For example, academic misconduct regulations often struggle to adequately define the unauthorised and undeclared use of generative AI by students and consequent outcomes. Institutions might have amended their definition of commissioning or plagiarism to address this, but the use of AI tools may not entirely fit within either definition.

ENAI working group members, including one author of this blog, have arrived at another solution - a new definition of ‘unauthorised content generation’ to avoid confusion in the natural language associated with commissioning and plagiarism. The new definition has been added to the ENAI Glossary and an associated exploration of the ethical use of generative AI by ENAI has been published recently in the International Journal for Educational Integrity.

Academic integrity and inclusivity

In her keynote, Dr Mary Davis of Oxford Brookes University, emphasised that academic integrity should be accessible and inclusive for all. Examples included creating an educational route through misconduct procedures and transparency in their outcomes to enable learners to learn and grow from their mistakes. She noted concerns around the over-representation of particular groups in academic misconduct cases including global ethnic minority, first generation, disabled and international students. She outlined ways of addressing this such as embedding academic integrity within the curriculum to create a climate that informs and promotes ethical behaviour and involving student ambassadors in creating an environment of positivity.

These ideas were explored through a QAA Collaborative Enhancement Project which Dr Davis led and which has produced extensive resources that are available now on the QAA website. The project resources link directly to the QAA Academic Integrity Charter.

Proofreading and ethics

Differences in proofreading expectations and practices were highlighted by several presenters who questioned the fairness and transparency of institutional policies which varied significantly in content and expectation. This included noting that many were difficult to find, as they were often embedded in high level documents that did not always use accessible language. Examples included everything from banning third-party proofreading through to institutionally-approved proofreaders with contracts that managed expectations on what was/was not allowed.

Subject-level differences in expectations of what would be considered poor scholarship or even plagiarism were also highlighted with concomitant differences in their associated marking schemes. This included STEM subject students who may not be expected to write extended assignments until the projects/dissertation stage of their final year. This leads to differences in academic literacy expectations between subjects that could be considered unfair. More cases of academic misconduct are likely at earlier study levels in the Humanities, but greater impact to degree outcomes are possible in STEM if only high stakes work can be considered to have been plagiarised. This has implications for expectations placed on postgraduate research students and their supervisors for research thesis writing.

Contract cheating

Contract cheating was the subject of several sessions. The importance of recording misconduct cases properly was noted. There was also discussion around the extent to which contract cheating and AI can share similar weaknesses, for example bad contract writers and AI can both produce non-sequiturs and poorly critique the literature by using inappropriate reference sources.

One presenter showed how tracing contract cheating cases back to the admission stage highlighted a rogue agent who was falsifying qualifications for about 46 applicants. This talk clearly showed the importance of cross-matching academic misconduct cases and the necessity for providers’ systems to be agile, providing salient data to allow removal of those who would devalue degrees and put the public at risk. The presenters expressed their wish to hold a national database of agents found to have falsified entry qualifications. There is some precedent for this in Australia where a similar approach exists for medical practitioners.

Hot topics in academic integrity

Two expert panels focused on current hot topics in academic integrity. The UK panel featured five members of the QAA Academic Integrity Advisory Group (Crockett, Dann, Draper, Glendinning and Lancaster) who talked about the strengths of the UK academic integrity groups located in the Midlands, North and Scotland, Wales, and London and the South East. UK delegates were strongly encouraged to join these networks.

The groups share (suitably anonymised) information about threats, practice and solutions to support one another in upholding academic integrity in their individual institutions. Three of the group chairs are also members of the QAA Academic Integrity Advisory Group which enables the information to be shared more widely across the UK. Common themes included AI threats and solutions, ensuring that the nomenclature of academic integrity is consistent, and the need to include the student voice in devising solutions such as making university policies accessible and encouraging students to act as advocates for academic integrity.

The impact of legislation in England to ban contract cheating (Skills and Post-16 Education Act 2022) was explored. It was noted that in Australia and Ireland where similar laws had been passed, their respective quality agencies had been assigned coordinating responsibilities in managing actions arising from the legislation.

Publication and research

Professor Michael Draper delivered a keynote giving a comprehensive account of contributions made by the Council of Europe and others leading to the development of strategies and legislation relating to academic and research integrity across Europe.

One parallel session showcased studies undertaken into research integrity and the co-authorship of academic papers. The presenters had developed a publication ethics training module (available to download under a creative commons licence from the BRIDGE website) addressing co-authorship and contributorship in research. A clear distinction between practices in STEM subjects and those in the Humanities was identified. Researchers in the Humanities tend to act alone or with a small number of collaborators and co-authors whereas it was not uncommon in STEM subjects for papers to be co-authored by many people with some individuals being credited with outputs on a more regular basis compared to the timeline for outputs in the Humanities. Interdisciplinary working is common in STEM where teams of people with complementary technical and practical skills tend to work together on multifaceted problems. These different practices can have an impact on career advancement and there is a risk that different practices between subject areas might not be recognised by promotion panels.

A key takeaway from this session is the need for individuals to be clear at the outset about the nature and extent of their contribution and author status to avoid potential disagreements at publication stage.

Concluding thoughts

There has been a step change in complexity in the last year with the rise of generative AI tools, and this means academic integrity requires more resource. The authors call for institutions to recognise the value of academic integrity and its practitioners.

The overarching messages from being surrounded by experts from across the globe are:

(a)  academic and research integrity is a diverse research field in its own right, with many sub-fields and great potential for further in-depth research

(b)  the concerns about maintaining integrity are the same everywhere, but sharing research outcomes, ideas, resources and successes by networking locally, nationally and internationally can help us to address some of the challenge

(c)   the rapidly evolving educational landscape necessitates a flexible approach to ensure that institutional regulations, policies and guidelines on academic integrity are kept up to date, fit for purpose and accessible to students

(d)  action to address academic integrity needs to take into account accessibility (for example to ensure regulations are written in language students can understand) and equality and diversity to ensure students with particular characteristics are not over-represented in misconduct cases

(e)  Academic integrity is the responsibility of everyone, it underpins best academic and research practice, providing a framework for inclusive action and the focus must not be solely on student misconduct.