Leaked Google memo shows it ‘deeply committed’ to Assistant, bringing LLM power

Back at Google I/O in May, we heard all about AI. In fact, if you replay the keynote and drink every time you hear the term “AI,” you’d likely die before the end of it. One thing we didn’t hear much about at I/O, however, was Google Assistant.

What chance does Assistant even have in a world with ChatGPT, Bing Chat, and even Google’s own Bard? This realization has caused some people to question whether or not Assistant is doomed. Thankfully, that doesn’t appear to be the case now that we’ve seen a leaked memo on the matter (via Axios).

The memo — which was meant to be internal among the folks at Google who work on Assistant, Bard, and related teams — explains some restructuring happening at the company. Teams are being shuffled around with one express purpose: “to explore what a supercharged Assistant, powered by the latest LLM [large language model] technology, would look like.” Google says it is “deeply committed” to Assistant and knows it is the key to helping to satiate “people’s strong desire for assistive, conversational technology that can improve their lives.”

The memo also says the company is “starting with mobile” for this journey. It’s not clear what that means, but Assistant for Android already has one LLM-powered trick in the works: summarizing web pages, per 9to5Google. It’s likely a lot of the LLM-powered features on the way to Assistant will come to Android/iOS first.

Theoretically, these LLM smarts would also come to smart home products powered by Assistant, such as smart speakers and smart displays. However, the memo makes no mention of these systems.

Unfortunately, the memo does confirm that some team members will be eliminated. Google is giving those affected employees a 60-day period to find positions somewhere else in the company.