Study Guide Supremacy

If you spend any time among people in the AI space, you’ll hear a common refrain. Large language models (LLMs) like Google’s Gemini, Anthropic’s Claude, and OpenAI’s ChatGPT are handy tools for searching the internet, separating facts from the chaff, and delivering information objectively. Of course, we’ve already invented a system for dispatching neutral information, and that system is called journalism. Herein lies a serious irony. Many LLMs are trained on content produced by journalistic institutions, but do not adhere to journalistic standards when rehashing the material. An LLM might regurgitate verbatim text from the New York Times without proper citation, or combine the work of real reporters with hallucinated quotes and misleading language. That poses not only a research problem but a major existential threat to news outlets, whose financial health depends, at least in part, on ad revenue from Web traffic. LLMs divert attention from news websites, while profiting from their archives and years of newsroom labor, and, in most cases, do not pay them a dime.