Mercedes Considered A Human Avatar For Its AI, Decided Against It

Artificial intelligence was the star of the show at CES this year, especially when it came to cars. And one phrase that kept coming up to describe in-vehicle AI functions was “human-like”: meaning, a system you interact with in a way that feels natural, almost like you’re talking to another person. But how far should that go, exactly?

Mercedes-Benz unveiled a big AI push for its forthcoming cars equipped with the new MB.OS end-to-end software, but the German automaker has also decided where “too far” might be. 

The new voice-controlled AI in Mercedes’ cars takes the form of a glowing three-pointed star that shifts colors and animations depending on how you interact with it. But in an interview with InsideEVs and other media outlets at CES, Markus Schäfer, a Mercedes board member and the brand’s Chief Technology Officer, revealed that at one point they discussed giving the AI a human- or human-like avatar. 

They ultimately decided against it, Schäfer said.

“Once, we thought we were going to have a face looking at you,” he said. “We did lots of studies internally before we decided not to do it. Some people like it, but quite a lot more people dislike the human face. So we have a ‘star cloud’ that somehow represents something like a human being, but it’s not a face.” 

Ultimately, this was probably the right decision, because the human representations of things like AI systems and even metaverse avatars has been the subject of heated debate. There are notable upsides to this, including human faces being something we react to “naturally,” as this Harvard Business Review story explains:

Why do we ascribe a personality to what we know is an artificial construct? Because we can’t help but respond instinctively to anything that appears to be human. Research from neuroscience shows that our minds are attuned to and react emotionally to facial signals. That’s why most people prefer to communicate face-to-face rather than over the telephone. In the case of digital humans, we know that what we see on the screen is an artificial construct, but we still connect instinctively to it, and we do not have to be computer experts to interpret the facial signals and make the exchange work properly.

That makes sense. But this subject has been incredibly controversial too, especially as generative AI gets better and bleeds into ideas like dead celebrities being reanimated for new movies and TV shows. Another example happened just this week: though there are other factors here at work here, including sexism and an apparent reluctance to pay actual humans, Mahindra Racing’s Formula E team took heat for unveiling a female-presenting “AI Ambassador” for the brand.  

Plus who could forget Mark Zuckerberg’s disastrous Metaverse avatar from a few years ago, something he’s apparently still mad about? And, what kind of face would even be used to represent AI in a Mercedes car? A completely artificial person? Bertha Benz? Dieter Zetsche? (The mustache would’ve been cool, though.) 

Many other automakers are exploring AI functions in cars too, but at this point, few if any have gone the route of human representation for their software systems. Possibly the closest example is BMW’s Mini brand, which is implementing a new “digital character” in the form of a dog named Spike as its in-car AI assistant. And that sounds quite a bit less unsettling than interacting with a fake human every time you need to set the navigation to a charging station. 

Regardless, Mercedes’ in-car AI sounds promising. The three-star brand already makes one of the very best voice recognition systems out there, and this new system can do things like show you nearby restaurants when you say “I’m hungry.” And Mercedes adds that it acts with “empathy,” sensing your mood based on various factors and responding in kind. Maybe a virtual person doing that is a little too Blade Runner 2049 for comfort. 

Contact the author: [email protected]