On August 5, according to Nieman Lab, a record number of eight Pulitzer Prize winners and finalists voluntarily disclosed the use of artificial intelligence with the judging committee this year, including five final winners and three finalists, setting a new high since the introduction of AI reporting rules for the Pulitzer Prizes in 2024.
This year's participants showed a more frequent and in-depth use of generative AI and large language models (LLMs), with technology applications mainly focused on basic editorial tasks such as efficient search of large volumes of documents, data organization, and cross-verification.
The Wall Street Journal used an internal LLM to quickly summarize thousands of public documents about floods in Texas; the Star Tribune of Minnesota used ChatGPT to translate a shooter's diary written in pseudo-Cyrillic script and had it verified by experts; The Associated Press used an LLM to retrieve tens of thousands of leaked documents about surveillance technology; and The New York Times used GPT-5 to reverse-validate journalists' manual classification of cryptocurrency cases involving the SEC, showcasing a novel workflow of "AI verifying human work."
Regarding this, Marjorie Miller, head of the Pulitzer Prize administration, stated that the news industry has now reached a consensus that "AI will be here to stay" and can clearly define its reasonable application boundaries (such as data collection and retrieval) and禁区 (such as writing and editing news articles). Additionally, the Pulitzer Prize official confirmed that the AI usage reporting rules will be officially extended and expanded to entries in book categories starting next year.
