Ford patents voice assist with configurable personality

Voice-assist systems are commonplace in cars today, but Ford wants to give them personalities.

A patent application from the automaker published by the United States Patent and Trademark Office (USPTO) on Jan. 2, 2024 (and originally filed on Aug. 17, 2020), discusses using machine learning to create in-car digital assistants with distinct personalities selectable by the user.

These personalities would be conveyed by the vocabulary used by the digital assistant when responding to prompts, as well as speech tone, pitch, and pronunciation. The vocabulary would be chosen by feeding the software examples of speech from people assumed to have the personality traits the digital assistant is trying to mimic.

Celebrities would be among the models for the digital assistant. Ford doesn’t appear to be suggesting that digital assistants would impersonate specific celebrities (which would likely bring some legal trouble) but it discusses using their social media posts, audio and video interviews, and books or articles about a celebrity with a given personality type to help train the digital-assistant software.

Ford digital assistant personality generator patent image

Ford digital assistant personality generator patent image

Ford digital assistant personality generator patent image

Ford digital assistant personality generator patent image

Ford also discusses giving the digital assistants names, hypothetical ages and genders, and even backstories. All so that drivers will get a more lively response when asking for directions to the nearest Starbucks.

As with all patent applications, it’s unclear if the ideas discussed here will make it to production. But the rise of ChatGPT means this is definitely on-trend.

At CES 2024, Mercedes-Benz unveiled an AI-powered next-gen user interface, while Volkswagen announced that it will integrate ChatGPT with its infotainment systems. So automakers do seem ro believe it will be better if we talk to our cars more.

Source link