The Washington Post’s new climate-focused AI chatbot is still working out its kinks. Futurism reported that the in-beta tool, which was announced on Tuesday and is meant to answer questions based on the paper’s archive of climate stories, refused to produce answers to multiple questions about AI’s impact on the environment. That’s despite the stories in question centering largely around AI’s environmental impact. “Sorry, but we could not generate an answer to your question,” the AI responded to questions on what AI’s environmental impact and carbon footprint are. “Please consult the articles below which may be relevant to your query.” The story then produced zero stories related to the questions, instead regurgitating the Post’s press release on the chatbot itself. The paper told Futurism the chatbot could only produce stories from its Climate & Environment and Weather verticals—though it later offered up a story from the Business vertical when asked a different AI-related question, which it did answer. The Post didn’t explain the discrepancy.
Read it at Futurism
