Subscriber OnlyOpinion

Reformed Leaving Cert would allow too many opportunities to cheat using AI

Every step is already susceptible to AI intervention, from brainstorming ideas to generating a structure for each section, to writing the entire piece

Teachers protesting outside Coolmine Community School in Blanchardstown last year over the implementation of the senior cycle redevelopment. Photograph: Bryan O’Brien
Teachers protesting outside Coolmine Community School in Blanchardstown last year over the implementation of the senior cycle redevelopment. Photograph: Bryan O’Brien

A recent report from Mary Immaculate College, University College Cork and the University of Limerick confirmed something maths teachers are tired of saying: the 25 bonus points for Leaving Cert maths have led to an overall decline in standards and an increase in student stress.

When the measure was mooted in 2010, people such as Prof Kathleen Lynch predicted exactly this outcome. It has taken 15 years to confirm that well-intentioned changes designed to improve matters harmed students and standards instead. The unintended consequences of the 25 bonus points, along with an unsatisfactory junior cycle reform, raised the stakes still higher for Leaving Cert reform.

The TUI has voted in favour of implementing senior cycle reform.

The shock is that the ASTI did not follow suit. Given that everything from pay rises to additional middle management promotions were dependent on voting yes, this is a courageous move which now gives the Minister a clear message about how principled the opposition is.

READ MORE

After years of dithering, in September 2023, accelerated reform was announced in a completely unachievable time frame.

As a consequence, reformed English, due to start in 2026, will now be delayed until the State Examinations Commission can figure out how to deliver an oral exam for every single Leaving Cert student. Yet, new Leaving Cert science subjects will begin from August, despite significant health and safety implications for all students carrying out individual investigations in inadequate laboratory spaces.

Well-off students with better facilities will easily outpace students from deprived areas.

Coursework counting for 40 per cent is a particular bone of contention in the age of generative AI. The proposed AI taskforce to be set up in August is not only too late for the incoming subjects, it is reprehensibly late for existing Leaving Cert coursework.

The general approach to coursework, aside from some subjects such as music and art, is to conduct an investigation, carry out a practical, or write an extended essay and then submit an account of it in a standardised booklet. This approach will continue in many of the reformed subjects.

The standardised booklets have a similar format: state the title, describe the approach, give the findings and comment on what you learned from the process.

Every step is already susceptible to AI intervention, from brainstorming ideas and generating a structure for each section to writing the entire piece. Even without plagiarism, this approach most benefits those with fluent language skills. In this newspaper, science teacher Adrienne Healy queried the value of writing a narrative about conducting a science experiment. As she says, asking musicians to demonstrate how they “play their guitar through a written report ... would be ridiculous and useless”. Why is it appropriate for science subjects?

External examination reinforces accountability. Before agricultural science was redeveloped in 2018, students carried out an investigation and wrote about it. An external inspector then asked questions of random students and examined their folders. This approach has now been dropped.

At junior cycle, in home economics 50 per cent of the marks are for a practical cooking demonstration that is externally examined, but at Leaving Cert, research followed by a practical demonstration is submitted via a journal. The practical is not examined.

The English oral is less susceptible to plagiarism, yet its introduction is being delayed while the science subjects are forced to go ahead

Leaving Cert reform presented an opportunity to test a range of skills other than writing, particularly vital in an age generative AI. As Australian researchers from Deakin University say bluntly, “GenAI challenges assessment validity by enabling students to complete tasks without demonstrating genuine capability.”

Regarding GenAI, the researchers contrast discursive changes versus structural changes in assessment at third level. Most changes have been discursive. Students are given guidelines akin to traffic lights as to how much they can use AI, red being not at all and green allowing usage at every stage.

This relies completely on student compliance. Discursive changes “direct behaviour they cannot monitor. They prohibit actions they cannot detect.”

In contrast, structural changes are “modifications that directly alter the nature, format or mechanics of how a task must be completed, such that the success of these changes is not reliant on” student compliance.

Ironically, the English oral represents a structural change because it tests essential skills in a way less susceptible to plagiarism. Yet its introduction is being delayed while the science subjects are forced to go ahead.

Structural assessment changes involve more human input, not less, including strategies such as oral exams and external inspection, all of which cost money. The Department of Education is already under pressure to cut costs. But what are the consequences for the future of education? Or student stress?

The current timetable for submission of coursework runs from November to May of sixth year. When there is coursework worth 40 per cent in every single subject, inevitably deadlines will have to move back into fifth year. This move to fifth year is already being resisted by the association of English teachers, Inote, because students will not be ready.

If submission deadlines move back to fifth year, or even if they are packed into sixth year, the result will be chronic stress with little respite, exactly the opposite of the planned outcome. Let’s hope it does not take 15 years to admit it.