While a lot of us might be comfortable with the voice controls in modern cars, they often have a bit of a learning curve as you figure out the right commands to make stuff happen. That's not the case with human assistants that can understand implied context, but they're (1) expensive and (2) slow to respond. The latter (humans) are what has been the force behind GM's OnStar service, where a push of a button can summon the voice of a human to help with everything from emergency assistance to finding the nearest sushi restaurant.
Except now, pressing that button could get you a robot. General Motors partnered with Google to run the "OnStar Interactive Virtual Assistant" (IVA) off of Google's Cloud Dialogflow AI tool. In fact, they've actually been doing it since last year, with the AI voice able to handle common non-emergency customer requests. This is a big deal for GM, as it means they'll have to hire, train, and employ fewer humans to staff OnStar call centers. For customers, it means faster responses and a less robotic voice than many in-car voice controls.
GM says the OnStar IVA is "now handling more than 1 million customer inquiries a month in the U.S. and Canada", which honestly is far more activity than I expected OnStar to have these days and explains GM's eagerness to implement AI in this fashion — handling 1 million customer inquiries a month requires a lot of humans, or just one powerful AI server farm. And as this is all handled server-side, the OnStar IVA is available in GM cars dating back to 2015.
For emergency calls, OnStar will still route drivers to human call takers.
While GM offers varying trial periods of OnStar on purchase of a new vehicle, the service does cost a minimum of $29.99/month, though plans starting at $44.99/month do offer unlimited in-car Wi-Fi. Then again, while I was surprised to hear that OnStar fields a million questions a month from drivers, I also am looking at my phone right now that has a pretty decent voice assistant built into it as well.
Implementing AI in this manner is unsurprising. OnStar employees only a several hundred people, but they contract out a lot of the grunt work for to call center contractors (as do many companies big and small). Implementing this AI-powered OnStar IVA not only gives OnStar more control over the system but it means they won't have to contract as many human call takers to field those millions of annual requests. It's a great thing for OnStar and GM, but it's an open question what this means for humans in the long run.
Read more
US automakers team up to install 30,000 EV fast chargers and rival Tesla
A coalition of American car manufacturers has formed an alliance to build a charging network to rival Tesla's Superchargers, and that's good news for everyone.
Google's AI search is simply too slow and too wrong to be useful
Google added a generative AI beta to its search results, and to be frank… it's too slow. Some of that comes down to the very nature of generative AI, but it also speaks to the current limitations of AI and the expectations of search.
Mercedes is putting ChatGPT into its cars for some godawful reason
There's a time and place for artificial intelligence, and in these early years that line is blurry. Sometimes we need to try implementing it somewhere to see if it's a good idea. Other times… it's really obvious.
The post-truth political campaign is officially here, thanks to AI
Political campaigns are no stranger to mudslinging, exaggerations, and outright lies. And now political operatives are enlisting generative AI tools to further their goals, the truth be damned.