This Week's News: Can AI Replace Teachers?
•
12 min read
This article has been machine-translated into English. Therefore, there may be nuances or errors in the content. The Danish version is always up-to-date and accurate. Read the original Danish article here.
Here is this week’s news roundup on artificial intelligence in the education sector. We have gathered the most important stories so you can easily keep up with developments.
This week you can, among other things, read about:
- A report commissioned by Dansk Arbejdsgiverforening (DA) (the Confederation of Danish Employers), which estimates that as many as 9,000 teaching posts could potentially be cut in 2040 through the effective use of generative AI.
- OpenAI launches a new voice mode, Advanced Voice Mode (AVM), which makes it possible to hold conversations with ChatGPT.
- Gender differences in the use of AI tools among Norwegian students.
- A new study shows that artificial intelligence can generate more original research ideas than humans.
- A dilemma around how to handle ChatGPT at universities. Research from Aarhus Universitet (Aarhus University) examines how AI affects teaching and academic culture.
- Workshops from Styrelsen for Undervisning og Kvalitet (the Danish Agency for Education and Quality) on the future of teaching and exams in light of generative artificial intelligence.
- New article on the use of generative AI among upper secondary school (gymnasium) students in social studies.
- Uddannelsesdebatten 2024 focused on intelligent education and the role of AI.
Happy reading
Artificial intelligence could revolutionise teachers’ work towards 2040
A new report commissioned by Dansk Arbejdsgiverforening (DA) points to artificial intelligence having a significant impact on the public sector in the coming years, particularly in education. The report, prepared by Boston Consulting Group, estimates that as many as 9,000 teaching posts — equivalent to almost one in four teachers — could potentially be cut by 2040 through the effective use of generative AI.
We have reviewed their report and written a commentary on their claims:
DA with exaggerated promises about generative AI in education
After reading Dansk Arbejdsgiverforening’s 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 as many as 9,000 teaching positions — equivalent to almost one in four teachers — could potentially
Per Størup Lauridsen
At the same time, we would like to refer to Svend Brinkmann’s highly relevant comment:
OpenAI releases its Advanced Voice Mode - not in Europe
OpenAI has introduced a new voice mode, Advanced Voice Mode (AVM), which makes it possible to hold conversations with ChatGPT. Although AVM does not make ChatGPT better, it changes the way we interact with our devices and creates an experience that feels more like talking to a friend than giving commands to a machine. Something you will particularly notice is that it is quite good at Danish, and there is very little waiting time before a reply arrives. Compared with current assistants such as Siri or Alexa, ChatGPT’s voice mode stands out by being able to conduct more advanced conversations.
Critics, however, point to the potential risks of this technology, as it can create a false sense of human connection. AI assistants like AVM can become an extension of social media that exploit our need for relationships, but without genuine human interaction. There are also fears that such technologies could lead to a dependence on “intelligent” devices that constantly seek to hold our attention.
OpenAI’s Advanced Voice Mode is not yet available in Denmark, but by using a VPN to US servers you can try the new voice mode if you have a paid ChatGPT subscription.
Advanced Voice Model - Google Zoeken
OpenAI’s new voice mode threw me into the uncanny valley
OpenAI’s new Advanced Voice Mode, reminiscent of the movie Her, sounds almost too human.
Gender differences in the use of AI among Norwegian students
Norges Handelshøyskole (NHH) (Norwegian School of Economics) has examined the use of AI tools among 595 male and female students. The study shows that 85% of male students use AI tools such as ChatGPT occasionally or always, while this is true for only 61% of female students. The researchers believe one reason may be that top-performing female students are more inclined to follow rules strictly, especially when the use of AI tools is not explicitly permitted. The researchers therefore propose more instruction in AI use to promote equal opportunities in the labour market. The aim is to harness the potential of the technology while developing all students’ competences in AI, as these skills are becoming increasingly important to future employers.
Female students use AI less than male students, and that is worrying
The best-performing female students opt out of AI tools, while men acquire AI skills, shows a new study from the Norwegian School of Economics. — I recognise myself in this, says student Jeanette Tisløv.
Solveig Mikkelsen
AI models show greater originality in research ideas than researchers
A new study shows that an artificial intelligence (AI) can generate more original research ideas than humans, at least within the parameters the study set. Using Anthropic’s Claude 3.5, the researchers behind the study generated 4,000 ideas based on scientific literature. At the same time, they asked 50 researchers to develop their own ideas, which were then evaluated by 79 independent reviewers. The reviewers did not know whether an idea came from humans or AI, and they rated AI-generated ideas as both more original and more exciting than those created by humans. However, the human ideas were rated as more realistic and practically feasible.
The results show the potential of AI models to find new research angles, but the researchers behind the study point to several limitations. For example, the human participants were asked to devise ideas in a short time, and the language of all submissions was edited by another AI, which may have affected the assessment of originality. Moreover, the question of whether AI models can surpass the most groundbreaking human ideas remains open.
Read more here:
Do AI models produce more original ideas than researchers?
The concepts were judged by reviewers. They were not told who or what had created them.
Gemma Conroy
Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers
Recent advancements in large language models (LLMs) have sparked optimism about their potential to accelerate scientific discovery, with a growing number of works proposing research agents that autonomously generate and validate new ideas. Despite this, no evaluations have shown that LLM systems can take the very first step of producing novel, expert-level ideas, let alone perform the entire research process. We address this by establishing an experimental design that evaluates research idea generation while controlling for confounders and performs the first head-to-head comparison between expert NLP researchers and an LLM ideation agent. By recruiting over 100 NLP researchers to write novel ideas and blind reviews of both LLM and human ideas, we obtain the first statistically significant conclusion on current LLM capabilities for research ideation: we find LLM-generated ideas are judged as more novel (p < 0.05) than human expert ideas while being judged slightly weaker on feasibility. Studying our agent baselines closely, we identify open problems in building and evaluating research agents, including failures of LLM self-evaluation and their lack of diversity in generation. Finally, we acknowledge that human judgements of novelty can be difficult, even by experts, and propose an end-to-end study design which recruits researchers to execute these ideas into full projects, enabling us to study whether these novelty and feasibility judgements result in meaningful differences in research outcome.
Chenglei Si
Universities in a dilemma: How should ChatGPT be handled?
New research led by Ida Schrøder from Aarhus Universitet (Aarhus University) examines how these tools affect both teaching and the broader academic culture. A study of more than 200 million assignments shows that 11% were partly written by AI, prompting many universities to consider whether to ban or embrace the technology.
According to Schrøder, there are two main approaches: some view AI as a tool for cheating, while others see its potential as a pedagogical resource. In her research, Schrøder has found that many universities struggle to find a coherent strategy, and responsibility for regulation is often left to faculties or individual lecturers. There is also growing concern about how AI contributes to technological stress and inequality among students, as well as how it affects climate responsibility at universities.
Universities therefore need to rethink their approach to teaching, and perhaps turn towards more participatory forms of assessment. At the same time, Schrøder points to a gender difference in approaches to AI, with younger men dominating the field, which may skew decision-making processes.
Should universities ban or embrace ChatGPT? New research explores the dilemma - Algoritmer
Generative AI has turned the university world upside down.
mathildeodamm.dk
Workshops on the future of teaching and examinations
Styrelsen for Undervisning og Kvalitet (the Danish Agency for Education and Quality, STUK) invites you to a series of workshops where you have the opportunity to contribute to how teaching and examinations in your subject can be developed and improved.
The purpose of the workshops is to involve upper secondary school (gymnasium) teachers in the development of teaching and examinations in light of the emergence of generative artificial intelligence, as well as to identify experiences, challenges and potentials in the use of generative artificial intelligence. As a participant, you will also have the opportunity to develop ideas for concrete assessment criteria and to ensure a meaningful alignment between teaching and examinations in the subject.
Please note that the workshops will not necessarily result in pilots or changes to the subject, but serve as a platform for professional discussion of teaching and examinations in the subjects. The input that emerges will be included in STUK's further work.
- Major Written Assignment, hf – 4 November
- Mathematics B, htx – 4 November
- Biology A, stx – 4 November
- SOP/SRP (stx, htx and htx) – 5 November
- Culture and Social Studies Group, hf - 25 November
- International Economics, hhx - 25 November
- Physics C, stx - 25 November
- English A (stx) – 2 December
- Danish A (stx, hf, hhx and htx) – 2 December
- Social Science A (stx) – 3 December
It is free to take part in the workshop, and we will provide catering and also cover your travel expenses. All workshops are held at Scandic Sydhavnen in Copenhagen, which is within walking distance of Sydhavn Station, and there is good parking on site. Places are limited for each workshop. Places are allocated on a first-come, first-served basis.
Sign up via these links:
Workshops on the future of teaching and examinations
Workshops on the future of examinations
Generative AI in Social Science
In a new article, Alex Young Pedersen examines the use of generative AI among upper secondary school (gymnasium) students in social science. The study shows that many students already integrate tools such as ChatGPT into their learning processes, and Alex Young Pedersen argues that schools should acknowledge this development and incorporate AI into teaching in a reflective way. He suggests using vignettes (short cases) and "show, don't tell" approaches to promote critical thinking about AI's role in education. The aim is to harness the technology's potential while developing students' independent, subject-specific competences.
Alex Young Pedersen Artificial intelligence and students' digital competences | POLIS – tidsskrift for samfundsfagsdidaktik
Authors Alex Young Pedersen
Uddannelsesdebatten 2024 - The people's meeting on intelligent education
Uddannelsesdebatten focuses each year on current questions of civic formation and education and was held on Thursday and Friday last week. The school town of Nørre Nissum provides the historical and authentic setting for the free people's meeting, which places strong emphasis on public enlightenment and democratic education.
Claus and I went along and came home very inspired. The focus was on intelligent education, and in particular on how artificial intelligence affects schools.
We won't write a long report, but would like to refer you to Anders Peter Nielsen's fine series from Uddannelsesdebatten, which appeared on LinkedIn in three parts:
What took place in the main hall was recorded, and below you can watch the livestream from the two days.
Other news this week
Expertise Not Included: One of the Biggest Problems with AI in Education
The debate about the effectiveness of AI chatbots in education continues amidst rapid AI advancements. Despite the broad knowledge and potential of AI systems, learners need expertise to effectivel…
Leon Furze
A regrettable and uninspiring response about digital learning experiences
The relationship between teacher and student may well be dominated by digital meetings, but a well-functioning relationship goes much deeper than chance encounters with bus drivers and hotel receptionists, writes Arve Hjelseth in a reply to Nedim Olsson Kasumacic.
Arve Hjelseth
The teacher is here to stay, but might the role change?
Debate: AI can streamline certain tasks and generate tailored plans. But the relationship between teacher and pupil is still a crucial point for good teaching.
Martin Johannessen
Generative AI In Classrooms: Cheating Or The Future Of Education?
Is generative AI in classrooms cheating or the future of education? Discover how AI can enhance learning and prepare students for the challenges of tomorrow’s workforce.
Nirit Cohen
“It is not going to reform teaching” | Alinea
When the mathematics teacher Jonas Mølgaard Lilholt prepares his lessons, he uses generative AI for inspiration, but also to take care of the routine work. At the same time, he incorporates artificial intelligence into his teaching to show pupils how generative AI can be a support tool – and not just a cheating tool.
Written by Angelica Petersen
How can AI be used in schools?
ChatGPT should not write your German essay for you. But what should it be used for instead?
See Sine's profile
At Slotshaven Gymnasium (upper secondary school), pupils are to get ”points” on their driving licence
Text_ Ida Abildtrup Photo_ Ida Abildtrup
New AI initiative from Vallensbæk Kommune (Vallensbæk Municipality) - IT Torvet
New guidelines are intended to ensure the responsible use of AI, including the safe use of services, data protection and ethical responsibility.
Malte von Sehested
Hilfe! Upper secondary school (gymnasium) teachers sound the alarm: We are on the brink of a collapse in language skills
Businesses are demanding employees with strong language skills, but the desire to study German, French and Spanish has fallen so markedly that we will soon lack teachers who can teach future generations.
Johannes Baunsgaard jobauerhvervplus.dk
Warning to Copilot users: Compromising a single employee can prove costly if you are not careful
Interview: Compromising a single employee can turn into an unpleasant affair if the security around the many AI assistants is not properly managed.
Kristoffer Ingemand Petersen Enterprise Editor
The danger is not the technology. The danger is polarisation among pupils
Getting school staff to embrace AI with equal parts critical reflection and professional curiosity is school leaders’ foremost task. A key point of attention is to stay focused on bringing all pupils along – regardless of their starting points. So says Professor Ole Sejer Iversen.
By Mona Samir Sørensen
Written by
MSc in IT | Co-founder of Viden.AI | Head of IT and Communications at TietgenSkolen
