The Washington Post recently unveiled ‘Climate Answers’ – an experiment that leverages artificial intelligence to help users discover and explore the agency’s climate reporting. Users’ questions are addressed directly from articles published by the Post since 2016.
“If the tool does not readily find a relevant article (even if we have published one), it won’t serve a reply: we set a high bar in our scoring of relevant articles for a given question,” noted chief technology officer Vineet Khosla in a blog post.
The AI chatbot finds information from the publication’s extensive archive on the subject to compose its response in everyday language, besides links to the pieces it is using as sources. It has some strict guardrails on it to prevent hallucinations or misinformation.
Post’s larger plans for integrating AI into its operations
It is important to ensure accurate and accessible climate information. The effects of climate change are impossible to ignore for anyone paying a slight bit of attention. But people need to become more aware of the crisis and the need for climate action.
The chatbot is a major step in the Washington Post’s larger plans for integrating artificial intelligence into its operations as part of its ‘Build It’ programme. The publication already employs a synthetic voice created by AI to read newsletters out loud.
“Over the past six months, the Product, Design and, Engineering teams worked closely with the climate team to build this new conversational interface to make it easy for readers to get answers to their climate questions,” Khosla added.
But there is a bit of irony involved
Google recently released its 2024 Environmental Report to highlight that its greenhouse gas emissions last year were 48% more than in 2019. It blames the increasing amount of energy required by its data centres, driven by the explosive expansion of AI.
Like other big tech firms across the globe, Google has gone all-in on investing in AI. The company has integrated its Gemini generative AI technology into some of its most significant projects, including Search and Google Assistant.
But AI has a major problem. It needs a huge amount of energy-hungry data centres or massive collections of computer servers. A generative AI system, such as ChatGPT, is likely to use around 33 times more energy than machines running task-specific software.
READ MORE : Can AI actually help solve the climate crisis? Some experts challenge the idea
