Google explains how Android Auto AI message summaries work

With the Samsung Galaxy S24 launch, Google announced that Android Auto would “automatically summarize long texts or busy group chats.” While we wait for this Google Assistant feature to launch, here’s more about how the AI message summaries will work on Android Auto.

Of note to us is how AI message summaries are officially branded as a Google Assistant feature. While all voice output on Android Auto is powered by Assistant, the announcement in January just credited “Android Auto” as being responsible for this capability. 

Additionally, while the support document just talks about this feature being available on Android Auto, it’s somewhat interesting how Google at the beginning says “you can receive message summaries of long messages on your Assistant-enabled Android device” (emphasis ours).  There’s nothing stopping this feature from being available on other audio-heavy form factors, like headphones. 

How to opt-in to Android Auto message summaries 

There are two ways to opt-in to this feature. The manual route involves going into Android Auto Settings > Notifications > Play AI message summaries. This on/off toggle is already available, but we haven’t been able to trigger AI summaries in our testing.

The other method involves Google prompting you the first time it’s available. Simply say “Yes” to give “approval to summarize” or “No” to have Assistant play the message in full. You can enable it at a later date from Settings. 

What messages get AI summaries 

Google Assistant on Android Auto can summarize “one long message or multiple messages from the same sender or the group.” Google defines a long message as being “over 40 words.”

Otherwise, the Assistant reads the message without summarization. If there are multiple conversations from different senders or apps, the Assistant reads them one by one.

The one example Google provided was a 3-line message that starts with “So about dinner tonight, I was thinking…” getting summarized as “Brianna asked if you wanted to have Thai for dinner and if you were still able to make it.”

AI Privacy

Google cautions that “there could be mistakes” given that “artificial intelligence generates your message summaries.”

On the privacy front, “Assistant doesn’t log messages or summaries,” while “interactions aren’t used to train the Large Language Model (LLM),” which is also the case for Magic Compose replies in Google Messages. Google doesn’t specify what model is being used.