DA’s exaggerated promises about generative AI in education

After reading Dansk Arbejdsgiverforening’s (the Confederation of Danish Employers, DA) report on the potential of generative AI in the public sector with a focus on education, there are some challenges we need to address.

The report, prepared by Boston Consulting Group, estimates that up to 9,000 teaching posts – equivalent to almost one in four teachers – could potentially be freed up in 2040 through the effective use of generative AI.

The report highlights five main areas where AI can support teachers’ work:

1. Preparation of teaching materials
2. Creation of interactive learning environments
3. Rapid assessment of assignments
4. Detection of cheating
5. Streamlining of documentation work

Danmarks Lærerforening (the Danish Union of Teachers, DLF) remains sceptical about the report’s conclusions. Regitze Flannov from DLF’s education committee points out that experiences from the COVID period have shown the importance of in-person communities, relationships and presence in teaching – elements that AI cannot replace. DLF does, however, see potential in AI as a supportive tool that can free up time so more teachers can be present in classrooms.

DA urges politicians to take an active stance on the implementation of AI in the public sector, including the Danish primary and lower secondary school (folkeskole). They recommend that requirements for AI implementation be incorporated into the annual economic agreements with municipalities and regions as well as in the Finance Act (national budget). At the same time, DA underlines the importance of involving teachers in this transition process to ensure that the technology helps create better workplaces with a focus on the core task and increased job satisfaction.

Challenges with the report

Fundamentally, I believe the report rests on questionable sources and expresses an overly optimistic view of generative AI’s role in teaching. This is particularly true regarding the assumption that up to 9,000 teaching posts can be freed up, which seems unrealistic without a thorough analysis of the pedagogical consequences. The report’s conclusions risk overlooking important factors such as pupils’ need for human interaction and the limitations of current technologies.

Below are my criticisms of the report and its conclusions:

1. The report does not mention the costs of implementing and operating these AI solutions, which are supposedly to replace teachers. It would be almost impossible to phase in AI technologies on the scale proposed. At the same time, experience shows that digitisation projects in the public sector often become more expensive and more complicated than first assumed (see sources).
2. It is assumed that the preparation of teaching materials, assessment and documentation can be made significantly more efficient, but it is unclear how these figures have been arrived at. Many of these tasks require human judgement, empathy, creativity and situational awareness. I believe the report underestimates and diminishes the professional expertise of the profession as a whole.
3. The report’s data basis relies primarily on Danish and international time studies of work tasks, supplemented with other, more or less disconnected, sources. I question whether this basis is robust enough to draw such far-reaching conclusions about the potential for reducing the number of teachers.
4. The report also fails to address the new ethical dilemmas and risks that extensive use of generative AI may entail. What about errors and bias in AI systems, which can have major consequences for pupils? If an AI model for assessing pupils’ performance is biased, it can lead to the systematic discrimination of certain groups. What about the risk of surveillance and loss of privacy? How do we ensure transparency and accountability in its use? What about continuing professional development for teachers?
5. Last but not least, the report does not address what any productivity gains from generative AI should be used for. Is it merely about cutting the public sector, or about freeing up resources to improve the quality of teaching? That political discussion is entirely absent from the report.

Overall, I am concerned that the report promises more than generative AI technology will realistically be able to deliver in the education sector within the foreseeable future. Artificial intelligence can, in time, support certain well-defined tasks, but the main focus should be to use any resources freed up to strengthen what only humans can do – presence, relationships, differentiated personal feedback, etc.

We should adopt a far more critical and nuanced approach to both the opportunities and the challenges of generative AI than is the case in the report from Dansk Arbejdsgivere. A more transparent, balanced and research-based approach to the potential is necessary before we draw hasty conclusions that could have far-reaching consequences for an entire sector.

Sources

DA: With artificial intelligence we can do without almost one in four teachers by 2040

AI could free up 9,000 teaching posts in 16 years, assesses Dansk Arbejdsgiverforening. No, artificial intelligence cannot replace teachers, but it may free up time so that there can be more teachers in the classroom, says Danmarks Lærerforening.

folkeskolenVillads Kirk Thomsen

DA: AI can free up 6,000 employees annually in the public sector

Artificial intelligence (AI) and its impact on the Danish labour market is hotly debated at the moment, and in that context Dansk Arbejdsgiverforening (DA) has commissioned a report from Boston Consulting Group (BCG) that looks at the opportunities associated with generative AI (GenAI) in the public sector towards 2040.

AI-PortalenMark Sinclair Fleeton

Generative AI brings great opportunities, but also challenges

In a very short time, artificial intelligence (AI) — including the rise of generative AI — has taken a place on the agenda of Danish top executives.

PwCPricewaterhouseCoopers

Public IT projects: An expensive affair with many disappointments

Reader debate by: Poul Rand, local chair of Liberal Alliance in Fredericia

DanmarkC TVMads Wiig Ehmsen

Five initiatives to fix the large public IT projects - IT-Branchen

There must be an end to large, problem-ridden IT projects in the state sector. Five initiatives from the government are to ensure that future projects are delivered successfully.

IT-BranchenRune Fick Hansen

An end to public IT failures? A new tool is to make digitalisation cheaper and more flexible

PROFESSOR: A new method for developing public IT systems makes it easier to digitalise complex workflows while also maintaining flexibility when reality changes. In this way we could prevent the many IT failures that have cost Denmark billions in recent years.

Altinget: Digital

https://digst.dk/media/dbmdbatx/analyse-af-barrierer-1.pdf