For years, higher education leadership dismissed anecdotal evidence of Large Language Model (LLM) bias as noise. That excuse is dead. New analytics from Kolaricon Consulting, utilizing Otterly.ai, provide hard data confirming that AI engines disproportionately recommend large public and nonprofit universities. The study evaluated over 30 unbranded search prompts across major AI platforms, including ChatGPT, Google AI Overviews, and Microsoft Copilot. The results show a stark concentration of visibility among a handful of institutions, fundamentally altering the enrollment funnel for online education.

The Duopoly of AI Recommendations

The data exposes a winner-take-most dynamic. Western Governors University (WGU) dominates the sample set, appearing in more than 60% of monitored responses. Southern New Hampshire University (SNHU) follows closely with nearly 50% brand coverage. Together, these two institutions account for more than half of the share of voice across the tested queries. The top four institutions in the sampleβ€”WGU, SNHU, and two other public/nonprofit entitiesβ€”command 80% of brand share. This concentration suggests that if a prospective student asks an AI for advice on online degrees, the algorithm has already narrowed the field before a human even reads the answer.

Positioning and the Bias Against For-Profits

It is not just about frequency; it is about prominence. WGU averages a position of 1.4 in AI-generated answers, while SNHU holds a position of 1.8. In contrast, Grand Canyon University, a former for-profit now operating as a nonprofit, languishes at an average position of 4.3. Liberty University also scores poorly on placement. This aligns with broader industry trends where publicly traded postsecondary companies have reported declining new starts, attributing part of the drop to AI search visibility. While Strategic Education’s CEO claimed no adverse impact from LLMs in Q3, the metrics suggest otherwise. The algorithm actively recommends alternatives to for-profits, creating a structural disadvantage for brands that lack the 'nonprofit' signal or the sheer scale of the public giants.

The Rise of Low-Traffic Citation Sources

Perhaps the most counterintuitive finding involves the sources LLMs cite. One might expect AI engines to pull from high-traffic domains like Reddit, YouTube, or Wikipedia. Instead, the study found that third-party education websites with relatively low traffic dominate citations. Research.com appeared in 23% of responses, Study.com in 22%, and Learn.org in 20%. These sites significantly out-cited official university websites, including WGU (17%) and Purdue Global (11%). Even EdSmart.org, which attracts only 57,400 monthly visits, appeared in 16% of responses. This indicates that AI engines are optimizing for structured, consensus-based data from niche aggregators rather than institutional marketing pages or user-generated content.

Key Takeaways

  • AI visibility is now a measurable metric, with WGU and SNHU holding over 50% combined share of voice in online education queries.
  • LLMs favor public and nonprofit institutions, actively biasing results against for-profit and Christian-affiliated schools.
  • Low-traffic third-party sites like Research.com and Study.com are the primary citation sources for AI recommendations, not major social platforms.
  • Traditional SEO metrics are obsolete for higher ed; leadership must track Brand Coverage, Share of Voice, and Average Brand Position.

The Bottom Line

The era of 'we aren't seeing anything' is over. If university boards aren't demanding data on AI search visibility, they are flying blind in a market where algorithms decide who gets discovered.