![](https://nordot-res.cloudinary.com/c_limit,w_800,f_auto,q_auto:eco/ch/images/1184185055634440559/origin_1.jpg)
The Washington Post has launched its own chatbot powered by artificial intelligence (AI) to answer its readers’ questions about climate.
Accessible online, the "experimental" tool’s "responses are based solely on published reporting by Post journalists," the publication said, adding that the outlet’s team has been covering climate and environment for more than two decades.
However, the tool only searches through articles from the past eight years.
Use of generative AI
The chatbot is powered by a large language model (LLM), a system trained on vast amounts of data to generate text in intelligible language for humans.
"Where in the US are sea levels rising the fastest?" "How does global warming affect biodiversity?" and "Can you recycle pizza boxes?" are among the suggested queries by the Washington Post.
In addition to the AI-generated answer, the Post links the stories relevant to the topic of the query nudging the users to use them to verify the information provided by the chatbot.
In its Q&A section, the staff indicated that the bot may not provide an answer when there is uncertainty in order to avoid hallucinations.
The outlet also said that the chatbot’s purpose wasn’t to replace journalists but to interact in a new way with the content that was already produced.
Generative AI in newsrooms
AI, especially generative AI (genAI), is being progressively implemented in newsrooms with over 70 per cent of people saying that they "used generative AI in some capacity," according to a survey published by the Associated Press (AP) in April.
When the first off-the-shelf tools like ChatGPT appeared on the market, a lot of publications quickly edited guidelines regarding their uses.
Several outlets negotiated deals with genAI companies.
For example, The Financial Times (FT) granted OpenAI access to its content to train its large language model in return for an undisclosed payment.
This deal followed the ones already signed by OpenAI with AP, the French title Le Monde, and the Spanish outlet El País among others.
However, the New York Times decided to legally fight the genAI companies. In December 2023, the Times filed a lawsuit against OpenAI and Microsoft, alleging that their language-training models rely on massive amounts of textual data coming from its articles.