// . //  Insights //  Evaluating Educational Assets In A Post-AI Landscape

Artificial intelligence (AI), and particularly generative AI, is rapidly changing the landscape of every industry. Things that we thought could not be automated or replaced are quickly becoming outdated. New opportunities are opening up that we did not think possible — including in the education and training sector. Investors evaluating assets within and around education may feel trepidation in making these investments as the potential risks AI poses to their investments remains uncertain.

However, there are also opportunities. Generative AI will disrupt the education and training industry, creating some winners and some losers. How investors leverage these opportunities will drive success or lead to failure.

From a business model perspective, assets that rely primarily on content creation, particularly when the stakes are lower (like homework or quiz prep), will likely have difficulty competing with AI. By contrast, assets that rely on models beyond solely content, including scope and sequence of pedagogy, standards alignment, or some level of human intervention, will likely be protected from this disruption.

From a market perspective, business models that are overly reliant on employment in sectors that could be disrupted by the technology are also at risk from reduced demand for workers and, ultimately, training to support those sectors if they themselves do not adapt to generative AI.

Key questions to evaluate the impact of generative AI on an educational asset

As investors seek out opportunities in education and training, there are several things to consider with respect to the risk and opportunity presented by generative AI. This checklist of questions lays out a framework for how investors should think about the potential upsides and downsides. In it, we use examples focused on test prep and study aids to easily contrast subsectors, but these questions are relevant for any investment opportunities within the education and training industry.

  1. How much does the business rely on content production alone?
    AI can create content cheaply and quickly. If the business focuses only on content production, it will at some point (likely in the investment period) be cannibalized by AI. For example, a company that only creates simple content like worksheets or flashcards will have a hard time competing against AI. Companies that add value beyond just creating content, such as editorial control or scope and sequence alignment or standards alignment, will fare better and can benefit from increased content creation efficiencies from AI.
  2. What are the stakes for the end user?
    For subsectors where 100% accuracy is paramount and/or for companies where the guarantee of a set standard or quality level is highly valued, AI poses a lower threat. AI alone cannot guarantee accuracy, and in fact is notoriously prone to “hallucinations,” where false information is passed on as truth. For example, in the legal profession, the bar examination is a high-stakes test with a high degree of difficulty. In addition, significant investment has been made by the candidate in the form of tuition, to be able to reach the final step of taking the exam. Students are unlikely to risk their career to inaccurate prep information to save what is a relatively low sum of money at the final gauntlet. In this and other high-stakes segments, test prep methods and content need to be highly accurate to meet the demands of learners; they will never rely on AI alone for their test preparation.
  3. Is there value beyond content?
    Business models that offer value beyond individual or standalone pieces of content will face less competition from AI. This creates a “moat” around the business model, protecting it from disruption. For example, the United States Medical Licensing Examination test prep provider UWorld not only delivers test prep, but also provides students with insights on where they rank compared with other users, which can help a student calibrate their need for further study. UWorld’s large customer base and back history of data will allow it to remain competitive even as a part of its core service model — the test prep question — could be disrupted by AI. AI can produce the content, but it cannot reproduce the rankings.
  4. What are the efficiencies to be gained using AI?For many companies, AI is not a threat, it’s a potential boon. There are productivity gains that AI may produce. Low-level work can be automated, creating more efficiencies within the business. Computer coding, for example, can be made more efficient using AI, reducing the need for staff. Companies that rely on software will receive the benefits of these efficiencies. Or, as another example, AI can write the first draft of a textbook chapter section, allowing an employee to spend more time refining and finalizing the chapter for higher levels of quality and to ensure accuracy.
  5. Is it dependent on other employment trends?These employment efficiencies can also be a risk for some businesses, particularly those that are sensitive to the demands of certain types of specialized employees. If AI makes Certified Public Accountant (CPAs) more efficient and the profession does not adapt in response, for example, firms will need fewer CPAs, and the demand for CPA test prep services will diminish.
  6. What level of human interaction is required?
    Anything that requires human service delivery will maintain its competitiveness within the market and potentially gain efficiencies from AI. Tutoring companies are the clearest example of these benefits: While content or practice questions can be produced by AI, the strategies and approaches for answering these questions will need to be delivered to students by a person to ensure effective learning. AI is unlikely to replace the teacher, professor or tutor, but it can help them be better and more efficient.

Answering these questions can help determine whether a potential investment will thrive and survive in a landscape dominated by AI.