This qualitative, exploratory study based on an analysis of results coming from conversational search engines, claims that smaller European Universities risk being increasingly less mentioned in the results of conversational search engines. For this study, we used different prompting techniques to trigger the conversational engines to come up with Master programs offered by the InnoEnergy university ecosystem, while never using the term ‘InnoEnergy’ as a contextual marker for those searches. We analyzed the output for emergence of better/less known universities.

When a specific concept is more digitally prevalent than a similar concept, the phenomenon is often referred to as ‘digital dominance’. The concept describes how established industry, highly visible universities or well-known brands tend to receive disproportionately higher visibility in search results, creating a self-reinforcing cycle. The effects can be explained by dynamics aligned with underlying bias, and recognition bias of the dominant form. As users are more likely to click on well-known universities in search results, increasing their click-through rates and further improving their key visibility within the LLMs. Bias is a known side effect of Large Language Models (LLMs) and was first seen as impacting digitally underrepresented groups, as well as depicting stereotypical reproductions of the information that was used to train the LLMs. But now we – the smaller, regional universities – are at risk of being gradually erased by the dynamics of the LLMs.

This blotting out of our universities is shaped by underlying algorithms that offer output, as well as by the actions of users of conversational search engines. There are multiple factors involved: higher historical traffic, more backlinks from reputable sources, longer established web presence, greater content volume, and better technical optimization. This combination of factors is growing with the increased use of conversational search engines, as these engines already trim down their initial results (in comparison to the ‘ranked pages’ option of traditional search engines).

Additionally, if new AI generated content is created referring to those better known universities, it adds to the content related to well-known universities, which reinforces the effect once again, which also trickles down into the traditional search engines. This means that more frequently used university brands, will become more prevalent in traditional search engines also, impacting the keyword effect on these search engines and in turn on the conversational search engines as well, adding to the digital dominance of the better known universities once more.

More research is needed to understand the full effect, and potential timeframes in which this digital dominance reaches a critical tipping point. In addition, we need to investigate what steps smaller universities can take to increase their digital visibility. In future research we need to consider how we can increase the digital visibility of our universities within these dynamic. It is our opinion that by strengthening our networks, and including industry and/or governmental partners to highlight what our universities and our students do in their marketing as well as in ours, can create a stronger digital presence across digital searches. This way avoiding that our universities will be erased by AI.

Keywords: AI, conversational search engines, GenAI, university, digital dominance, competitiveness.