Internet / Google

Google begins opening Bard to the public

Google's conversational AI experiment is reaching more users, with an explicit warning that fluent answers can still be wrong.

Original Google Bard experiment wordmark
Google / CC BY-SA 4.0

Google has begun opening access to Bard, its conversational AI experiment, initially in the United States and United Kingdom. The service offers a way to ask questions, explore ideas and request help with writing through a back-and-forth conversation.

The interface makes a familiar distinction harder to see. A search result points toward a source; a generated response can arrive as a complete paragraph that sounds settled. Google acknowledges that Bard can produce inaccurate information, so the form of an answer should not be mistaken for proof of its reliability.

For early users, the strongest approach is to treat the service as a starting point for exploration. Comparing suggestions, following up and checking important claims remain necessary parts of the task. The experiment is as much about learning an interaction style as testing a new website.

Source: Google announcement ↗

← GPT-4 accepts images and improves difficult language tasksVision Pro asks whether the next computer belongs on your face →