AIs ability to recognize patterns and hyper-personalize creates a large potential impact on education and mental health. Knowledge is power, and a pathway to development. Therefore, AI offers constructive opportunities for significantly enhanced learning and for mental well-being. But there are major ethical and political risks associated with this.
This blog post is the sixth in a series on AI’s disruption of innovation
This blog entry is the sixth in a series about the disruptive impact of a technology based on a new form of intelligence that is self-learning, universally enabling, and allows for deep customization:
- Where does AI impact today?
- What is new and disruptive about AI?
- AI’s effect on robotics and experimental automation
- AI’s effect on climate and meteorological research
- AIs impact on social sciences
- AIs impact on education and mental health
- AI’s impact on biotechnology and pharmaceuticals
- AI’s effect on materials research and quantum chemistry
- AI’s effect on theoretical physics and mathematics
- What should investment strategies in AI take into account?
Knowledge is power, and a pathway to development
Knowledge is power, and power is the biological ambition of all social beings: comparative advantage. But greater knowledge also increases overall opportunities for socioeconomic growth in knowledge economies. In addition, improved general well-being and mental health increase the likelihood that growth opportunities will actually be realized.
There is enormous commercial potential in the field of knowledge dissemination
The field of knowledge dissemination (education) accounts for between 6% and 10% of global GDP (USD 7–11 trillion). Yet the EdTech sector accounts for “only” about USD 160 billion and is expected to “only” double by 2030.
AI is particularly well-suited as a catalyst to unlock this untapped market potential. First, LLMs are good at generating clear and balanced answers to complex questions. In addition, LLMs can hyper-personalize their responses, based on personal psychological understanding as well as text analysis, response patterns, and more. At the same time, the marginal cost of deep personalization is close to zero.
Differentiated learning is the old dream now becoming reality …
Differentiated learning means personalized, tailored education. Students thus learn at precisely the individual level where they can keep up and feel motivated. It is expected to raise educational attainment when difficulty level and learning method are adjusted in real time to the student’s mood and preferred learning patterns. In addition, it leads to ongoing quality assurance, thereby reducing the need for formal examinations. This applies both to school education and to retraining and continuing education of adults. Here, agentic workflows can be connected to existing work functions, making the transition to further training seamless.
… and this is why attempts to realize the potential have been ongoing for a long time
Test platforms such as DreamBox, Carnegie Learning, Stanford’s Education Data Science Lab, MIT’s Teaching Systems Lab, and Khan Academy today use layers (tutors) built on top of ChatGPT-4. OpenAI’s strategy remains, for now, to stay at arm’s length from the market.
- AI in school education was valued at around USD 5 billion in 2024. Over the next 10 years, the amount is expected to grow to USD 112 billion.
- Learning AI should NOT make schoolteachers redundant. Instead, the role is shifting toward relationship-building and broader pedagogy, i.e. sparking curiosity and a desire for deeper engagement.
The risk elements involve deeper ethical questions such as cultural bias (e.g., whether AI is trained on Western-centric historical narratives) as well as surveillance and deep personal learning. The more deeply AI knows the individual, the greater the risk of AI-driven manipulation.
The risks are being actively debated in academic circles. Columbia University, among others, argues that universities should assume the role of moral guardians, responsible for setting boundaries on what should or should not be pursued, and how history ought to be described.
Mental health is relatively new but holds vast technological potential
Since Covid in particular, there has been a sharp rise in the number of people with mental health conditions, especially those related to loneliness and insecurity. This creates a very large commercial potential.
- The global market for behavioral health today stands at USD 127 billion. It is expected to grow to USD 173 billion over the next 10 years. The potential could be many times larger, but the field is ethically controversial. However, technological maturity should be able to address several of the key ethical challenges.
AI is considered especially well-suited as a supportive tool for therapists and in psychiatry. Here, analysis and diagnostic support of conversations (tone, word choice, etc.) can, for example, identify patterns of anxiety, depression, and stress.
- Several Big Tech companies, such as Microsoft, are highly focused on AI for mental health. Similarly, xAI’s Grok model has a development focus on “social connectivity.”
- Amazon and Microsoft have both researched conversational models for screening depression and PTSD. In addition, layers are currently being developed in a Jungian context, with symbolic mirroring, dream analysis, and more. These can draw the therapist’s attention to possible themes. Finally, both MIT and Stanford have developed AI models that measure empathy and evaluate the quality of therapy sessions.
Therefore, several products have already been launched with promising success
Among the promising digital therapists and conversational AI agents are Woebot (Stanford’s CBT-based model), Replika, Wysa, and Tess, all of which use therapeutic techniques. Wysa is widely used in many low- and middle-income countries, while Tess is more targeted at HR and healthcare systems. Replika aims at broader emotional support.
Finally, new GPT-based layers are being developed, targeted, for example, at lonely individuals. These layers may especially support elderly people as well as those processing grief.
The risks are tied especially to deeper ethical questions. For example, who is responsible if an AI layer provides harmful advice or reinforces cultural bias? Who should have access to the data (tech firms, insurance companies, etc.)?
The largest commercial potential may turn out to be the market for loneliness and grief support.