- Published on
OpenAI Most Often Named in U.K. Companies’ AI Disclosures, Study Finds
- Authors
- Name
- DeepHorus
A new study by European AI platform GreenPT has found that OpenAI is the most frequently named AI provider in public documents linking U.K. companies to artificial intelligence technology. The report, released in November, analyzed 1,377 documented links between British businesses and AI suppliers, with OpenAI accounting for 670 of them—nearly 49% of the total. The findings underscore OpenAI's dominant mindshare in the U.K. market, even as the study's authors caution that the data does not establish actual market share or adoption rates.
The research arrives amid growing corporate adoption of generative AI and increasing regulatory attention on how companies disclose their use of such tools. GreenPT, which markets privacy- and sustainability-focused AI services, reviewed publicly available materials including company statements, supplier case studies, and privacy notices. Each time a U.K. company was linked to an AI provider in these documents, it was counted as one reference. The study's main ranking placed OpenAI far ahead of rivals: Anthropic followed with 322 links (23.4%), a combined open-source category ranked third with 235 (17.1%), and Google trailed with 150 (10.9%). Together, OpenAI and Anthropic accounted for 72% of all documented links, highlighting the two companies' prominence in public AI disclosures.
However, the study's methodology includes important caveats. Because each company was counted once per provider, the 1,377 total does not represent 1,377 distinct businesses; a single firm naming both OpenAI and Anthropic would appear under both. More significantly, GreenPT estimated that about 31% of the links involving OpenAI, Anthropic, and Google lacked sufficient evidence of actual use. Some references involved proposed projects, outdated statements, conditional language, or pilots that may never have progressed beyond planning stages. To assess evidence quality, GreenPT examined a subset of 734 links across 15 providers. An entry was classified as "supported" only if it included disclosed use or an explicit supplier identification with a corroborated U.K. company connection. Under this standard, 69.9% of OpenAI links qualified, compared with 68.5% for Anthropic and 65.7% for Google—differences the researchers said were within statistical uncertainty.
The technical approach relied on document review rather than a representative survey, meaning companies that never publicly discussed their AI suppliers were excluded. As a result, the rankings reflect public visibility rather than actual deployment volumes or spending. Still, the study offers a rare quantitative snapshot of how AI providers are named in corporate disclosures. OpenAI's lead may partly reflect its first-mover advantage in generative AI and the widespread media coverage of its ChatGPT product, which has made it a default reference point for AI adoption.
The implications are twofold. In the short term, the findings reinforce OpenAI's brand leadership in the U.K. and could influence procurement decisions as more companies formalize their AI strategies. But the high proportion of uncorroborated links also signals a transparency gap: many firms may be eager to signal AI adoption without providing concrete evidence. This could invite scrutiny from regulators and investors who want clearer disclosure of AI risks and dependencies. Long term, as AI regulation tightens—particularly with the U.K. government's pro-innovation approach evolving—companies may face pressure to move from vague mentions to verifiable statements of use.
Industry reactions have been muted, but observers note that the study's limitations are typical of early-stage AI tracking. Competitors like Anthropic and Google may see the results as a benchmark and an opportunity to increase their visibility in U.K. corporate communications. The strong showing of open-source models across all nine industry groups examined is also notable, suggesting that cost and customization concerns are driving some firms away from proprietary APIs. Community discussions have focused on the need for standardized AI disclosure frameworks, similar to those for cloud providers.
Looking ahead, GreenPT's findings are likely to be updated as more companies disclose AI use and as methodological refinements are made. The study's authors have not announced follow-up research, but the approach could be extended to track changes over time or to compare with other markets. For now, the report serves as a reminder that public AI disclosures remain an imperfect proxy for actual adoption—and that the race for enterprise AI mindshare is far from over.
