Google was reportedly shifting its Assistant team toward Bard, its conversational AI, in a move that could bring chatbot-style exchanges to voice assistants. The idea would connect Bard’s language capabilities with Assistant’s speech recognition and its presence on phones and Nest speakers.
A reported change in focus
CNBC reported that Google was reorganizing the Assistant team around Bard, citing internal emails it had obtained. Sissie Hsiao, vice president and head of the Google Assistant business unit, said in a memo that the team should prioritize Bard technology going forward.
Hsiao described the work as a way to support and carry out the opportunities ahead, while emphasizing delivery that has an impact on users. The report suggests Google was considering how Bard’s chat technology might become part of Assistant, though it does not describe a finished product or a launch plan.
The shift reflects a broader possibility: a voice assistant could do more than respond to short, isolated commands if it could draw on the more flexible language abilities associated with chatbots. That could make speaking to a device feel more like an ongoing exchange.
Why speech and chat could fit together
Google had already shown early demonstrations of LaMDA, the AI model behind Bard, handling questions in scenarios resembling Assistant interactions. Pairing conversational language understanding with Assistant’s speech recognition could support more fluent spoken conversations with a computer.
The combination could also connect different ways of using devices. Someone might speak to Assistant on a smartphone or a Google Nest speaker, then continue through text chat on a PC or notebook. The source presents this kind of transition as a potential advantage, rather than a capability Google had already delivered.
Google’s existing distribution may matter in that contest. The report points to hundreds of millions of compatible Android smartphones with Assistant preinstalled as a large base of devices where an integrated experience could reach users.
The competitive pressure
ChatGPT and Microsoft were part of the competitive backdrop. Integrating Bard into Assistant could give Google a route to put conversational AI in front of people through products they already use, while connecting voice and text in a way that browser-based chat alone does not.
That potential advantage was not assured. The article also noted reports that OpenAI was working on a ChatGPT app that could include speech recognition and run on both iPhones and Android smartphones. If so, voice interaction would not necessarily remain a differentiator exclusive to Google.
The strategic question was therefore not simply whether a chatbot could answer questions. It was whether users could move naturally between devices and input methods, and whether that experience would be useful enough to make an assistant feel more capable in everyday life.
A separate dispute over Bard’s training data
The report also covered claims about how Bard was trained. The Information said Google researchers had allegedly taken ChatGPT-generated text from ShareGPT, a website where people shared conversations. Using such data for training would violate OpenAI’s terms of service.
Steven Tey, associated with ShareGPT, posted a claim that Google had relied heavily on the site’s data and connected it to the removal of ShareGPT’s Explore page, which he said held over 112K shared conversations. Those claims were disputed: Google spokesman Chris Pappas said Bard did not include data from either ChatGPT or ShareGPT.
The allegations and Google’s denial were distinct from the reported Assistant reorganization. Together, they showed two issues facing the company as it pursued conversational AI: how to bring the technology into products people already use, and questions about the material used to train it.