The Washington Post introduced a new AI-driven chatbot on Tuesday to address user queries about climate change. This tool, named Climate Answers, pulls information from Washington Post articles to provide accurate responses.
Currently, Climate Answers is designed to answer only climate-related questions. However, according to Chief Technology Officer Vineet Khosla, the Post plans to expand the chatbot's capabilities to cover more topics in the future.
Developed by the Post's product and engineering teams alongside its climate editorial team, Climate Answers leverages AI technologies from firms like OpenAI and Meta's Llama. This collaboration powers a large-language model that sources answers from various articles within the Post's extensive climate coverage.
The Washington Post is cautious about the information the chatbot provides. If the AI cannot confidently answer a query based on vetted post-reporting, it will refrain from responding. This approach ensures that the Post's journalism backs all chatbot answers. Users can report any misleading answers through a feedback form.
The Post has been increasingly experimenting with AI in its products. Last month, it launched an AI-driven article summary tool that generates concise summaries for about 10% of its stories. This feature is expected to expand to all articles eventually. Under the leadership of new CEO Will Lewis, efforts to integrate AI have accelerated.
Climate Answers has been in development for six months. Despite concerns that AI summaries and chatbot answers might reduce engagement with full articles, Khosla remains optimistic. Research indicates that younger readers prefer summaries over headlines to decide whether to read further, potentially boosting engagement among this demographic.
Unlike some peers, the Washington Post has not yet licensed its content to an AI firm. Khosla stated that the Post is open to partnerships that expand its journalism but emphasizes the importance of fairness in any such agreements.