The Washington Post has unveiled a new AI chatbot called “Ask The Post AI,” which aims to answer user questions accurately by relying on the newspaper’s content. The new AI chatbot follows the Climate Answers chatbot, which the publication released earlier this year. Unlike the climate journalism-focused Climate Answers, Ask the Post culls its answers from anything published since 2016.
The AI uses algorithmic ranking to match answers to questions and ensure they are relevant. Like Climate Answers, the new AI chatbot enforces strict guardrails on how it responds. If there aren’t any articles it defines as worth citing, it won’t answer the question at all. That way, it avoids any compulsion by the AI to hallucinate or provide wrong answers. The AI will just say it can’t answer at all.
Interactive AI Journalism
“This is the next chapter in building habits for our next generation of users,” Washington Post chief technology officer Vineet Khosla explained in a blog post. “The changed search experience across the industry calls for us to meet the moment and meet audiences how, when and where they want to be served with an updated user experience.”
“Ask The Post AI” arrives as The Washington Post has seen an uptick in interest from readers on its post-2016 coverage. The heightened reader engagement combined with new AI tools might serve to draw in and keep new readers as well. Along with Climate Answers, The Washington Post has been testing several other AI tools, including AI-generated audio recordings of news articles and AI-written summaries of articles.
There’s an obvious appeal for streamlined information access that supplies reliable answers without long research in the archives. Similar tools will likely crop up at the intersection of AI and news media elsewhere. For instance, Meta and Reuters have partnered to supply Meta AI with information from Reuters articles.