Credit rating agencies are facing the threat of AI, with their share prices taking a major hit this week following the release of new analytics tools by the ‘Claude’ AI services provider, Anthropic. Market sources have suggested to The DESK that AI-generated ratings are more appealing than traditional models, and academic research agrees.
A June 2025 study by academics from Harvard, Chicago Booth, NYU and Princeton noted that AI and machine learning could pick up on ratings downgrades, using signals from investors’ holdings. Credit spreads across bonds, from safer to riskier, were better explained using these tools than by models using traditional risk measures, the work found.
Deepak Khurana , a JP Morgan fixed income product manager, has suggested that traditional models used by credit ratings agencies are outdated, with the ratings they produce based on stale data and their methods opaque. AI could be used to ingest and process data faster and react to ongoing events, he said, with ratings updated more regularly.
“Will this kill rating agencies? Maybe not. But it will force them to evolve,” Khurana stated on LinkedIn.
So far, agencies have publicly focused their AI tools on the research and analysis element of the ratings process, rather than ratings generation.
Moody’s is already using generative AI in its credit analysis, stating that it offers “a new standard of accuracy, depth, and speed across credit risk analysis”.
S&P Global and Fitch Ratings launched AI tools within their research platforms last year. The former’s Credit Companion went live in May, and the latter’s Fitch Genie chat solution was introduced in July.
©Markets Media Europe 2025
