The Washington Post debuts AI chatbot
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Maura Losch/Axios
The Washington Post on Tuesday launched a new AI-driven chatbot on its site that answers user queries about climate with answers pulled from Washington Post articles.
Why it matters: For now, the tool is only built to answer user queries about climate. In the future, the Post plans to expand the chatbot to other topics, said chief technology officer Vineet Khosla.
Zoom in: Climate Answers, as the chatbot is being called, was developed by the Post's product and engineering teams in conjunction with its editorial team covering climate.
- The Post is working with several AI firms, including OpenAI and Meta's Llama, to power its own large-language model that surfaces answers to climate questions based on multiple articles across the Post's coverage of climate.
Between the lines: For now, the Post is treading carefully in what it feels comfortable surfacing to users.
- If the chatbot doesn't feel as though it can adequately answer a user query based on published and vetted Washington Post reporting, it will simply decline to provide an answer.
- The goal is to ensure that all answers "can be backed up with our journalism," Khosla said.
- Users can provide feedback about misleading answers to the Post via a feedback form.
Zoom out: The Post has been doing more to experiment with artificial intelligence in its products.
- Last month, it quietly launched a new article summary product that summarizes a given article using generative AI.
- The experiment, for now, is available on around 10% of the Post's stories, per Khosla. Eventually, the Post hopes to expand summaries to all of its articles.
The big picture: While the Post's focus on building new AI tools predates the arrival of its new CEO, Will Lewis' leadership has "accelerated" the effort, Khosla said.
- Climate Answers has been in the works for the past six months.
- Asked whether the Post was concerned about AI summaries and chatbot answers cannibalizing engagement with the Post's articles, Khosla said the company is "definitely watching the metrics" but isn't worried.
- Younger readers, he said, will often rely on story summaries over headlines to determine whether to read further, according to research his team conducted.
- That makes him hopeful that AI summaries could actually increase engagement with its content among younger cohorts.
What to watch: Unlike some of its peers, the Post has yet to strike a deal to license its content to an AI firm.
- Khosla said the Post "will talk to any company that helps us expand our journalism, but we also want it to be fair."
