When students in Denmark return to school this week after their summer recess, they will do so under a new regime governing how they can use generative artificial intelligence (AI) technologies for their assignments.
For example, the 9,000 students in upper classes (the rough equivalent of sixth form) who have to complete the Større skriftlig opgave (SSO), or major written paper, as part of their final assessments, will have to defend their essays.
Although not conceptually equivalent, SSOs are not entirely dissimilar to school-based assessments (SBAs), which the Caribbean Examinations Council (CXC) is phasing out for regional students, for the same reason the Danes are tightening oversight of the major project paper that their higher preparatory students typically work on during their final two years. Both are keen to maintain the integrity of their exams and the certification they provide and are concerned that students are increasingly using AI to cheat, especially in unsupervised assignments.
In addition to students having to defend their SSOs, the Danish education minister, Magnus Heunicke, has ordered schools to monitor the use of computers during written exams, including installing firewalls to limit the networks to which students have access. He has also advised schools to assign more written work during supervised class time, rather than having assignments done outside of school hours.
“Unfortunately, we have a problem with AI cheating in upper secondary schools,” Mr Heunicke said. “Action is needed now.”
He added: “I will engage in dialogue with schools, teachers and students about what can be done in the short and slightly longer term, so that we can ensure that AI does not undermine students’ skills, academic abilities and, above all, their ability to think for themselves.”
NOT UNIQUE
The challenge faced by Mr Heunicke is not unique to Denmark. Across Europe and elsewhere, education authorities are grappling with similar issues. While the technologies have vastly widened the environment for teaching and learning, they have similarly expanded opportunities for cheating.
How countries have responded to these challenges varies. For instance, in Denmark’s north-eastern neighbour, Sweden, the government engaged in a national conversation with education stakeholders on AI, attempting to determine, among other things, how significantly it is being used by students to cheat and if grading was being compromised by the deployment of the technologies across the system.
The Swedes concluded that they had a problem. Unlike the Danes, they didn’t opt for a mandatory centralised solution. Instead, the educational authorities in Stockholm authorised municipalities and principals to design their systems to tighten exams, but emphasised the return to more monitored, in-class, handwritten assignments.
France has been moving towards greater restrictions on digital devices during its annual continuing high school evaluations ( contrôle continu) as well as in the high-school qualification exam, the Baccalauréat. Additionally, the French have reduced the amount, and kinds, of assignments students get to take home, requiring that more work be in class, with limited access to the Internet.
In Britain, examination boards have tightened controls over the use of AI by students, including, in some cases, requiring that they provide written declarations confirming their authorship of unsupervised coursework. The rules also allow teachers to undertake ad hoc reviews of students’ work, including oral questioning, to compare the unsupervised coursework against students’ historic classroom performances.
There is no single response to the opportunities and challenges being thrown up by AI, although there are significant overlaps in the answers. In other words, while there can be overarching solutions to problems such as students using AI to cheat, regimes will have to be tailored and refined to suit national circumstances.
NATIONAL CONVERSATION
It is in that context, and with CXC having felt compelled to fully phase out SBAs by 2028, that The Gleaner’s Editorial Board again urges Jamaica’s education authorities to urgently convene a national conversation on a policy for the use of AI technologies in Jamaica’s education system. The Jamaica Teachers’ Association (JTA) has also called for such a dialogue.
There is no data on how much Jamaican students use AI to cheat. But, given the CXC’s action on SBAs, it is reasonable to assume that the examination body concluded that the practice is widespread across the Caribbean, including in Jamaica.
What can, however, be definitively said about Jamaica, given the findings of two University of Technology, Jamaica (UTech) researchers, Paul Golding and Tiou Clarke, is that AI is used by practically all students at the university and high schools, 100 per cent and 96 per cent, respectively. Additionally, 87 per cent of university teachers reported AI in professional tasks, although the specifics of the use are not clear.
AI is changing the world at a pace that would have been almost unimaginable not long ago. Yet, the technology remains in its infancy, although the consequence of every advance is exponential.
Societies, not least the education sector, have to embrace the opportunities offered by this, while guardrails are established to protect the integrity of the things whose sanctity must be cherished. These include certifying that individuals possess the knowledge they are supposed to have acquired after a course of study.
With respect to the education sector, the board agrees with the JTA, and the discussion must involve more than principals, teachers and school boards on an individual basis. The conversation demands all stakeholders.
