Google announces Bard

Google announced Bard, an “experimental conversational AI service” built on LaMDA, the dialogue model the company had been developing internally since 2021. The post said Google would initially release “a lightweight model version of LaMDA” that needed less computing power, opening it first to trusted testers before a wider public release “in the coming weeks.” CEO Sundar Pichai described it as combining “the breadth of the world’s knowledge with the power, intelligence and creativity” of Google’s large language models.

The announcement came ten weeks after ChatGPT’s launch and the day before a scheduled Microsoft event widely expected to put OpenAI’s technology into Bing search — timing read by most coverage as a defensive move to establish Google’s own position before Microsoft’s announcement landed.

The launch was quickly overshadowed. Google published a short promotional GIF of Bard answering a question about discoveries from the James Webb Space Telescope; the answer stated the telescope had taken the first pictures of a planet outside the Earth’s solar system, a claim that was false — the first such image had been taken years earlier by ground-based telescopes. Reuters flagged the error on 8 February, the same day Google held a related AI event in Paris, and Alphabet’s shares fell roughly 9%, erasing about $100 billion in market value in a single trading session.

The episode became a recurring reference point in coverage of the “chatbot arms race” that followed ChatGPT’s release: a factual error in a marketing asset, rather than in the product itself, briefly became a proxy for judging whether Google’s AI work was as far along as its rivals’. Bard itself did not open more broadly until Google put it in front of a public waitlist the following month.