Software-defined vehicles: is artificial intelligence the solution—or the problem?

While artificial intelligence (AI) promises to revolutionize mobility, the gap between digital hallucinations and physical reality suggests the industry faces a complex battle for software dominance that is only just beginning.

“AI can’t yet read maps,” said HERE Technologies CEO Mike Nefkens. “Maps are lines and polygons. AI reads text, code and other data features, but it can’t read lines and polygons. It hallucinates dramatically, and we have hundreds of examples of that.”

In a straight-talking interview with Automotive World, Nefkens shared what’s really going on with AI and software-defined vehicles (SDVs). The headlines make it sound like everything is already working smoothly, but out in the real world and on actual roads, it is messier than that.

According to Nefkens, the industry is in the middle of a big shift, moving away from patchwork, “Frankenstein” systems and toward software that is actually built to work together from the ground up.

Nefkens says a big part of the problem comes down to this. Teaching AI to work with text is one thing. Teaching it to understand the real world and physical space is a whole different challenge. He shared a few examples where AI even got basic navigation wrong, like putting famous landmarks such as London’s Big Ben in the wrong place. Mistakes like that show why reliable map data is still the backbone of future mobility, no matter how smart generative AI gets.

And the growing pains are not just about maps. The shift to SDVs has been rough for traditional automakers. Early efforts to “add software” to existing systems often created bloated, complicated setups that did not actually make things run better.

Nefkens calls this phase “SDV 0.5,” and contrasts it with the cleaner, software-first approach he sees from fast-moving Chinese startups, which he labels “SDV 1.0.”

The conversation also touches on the global race to innovate. Chinese carmakers are moving at what many people call “China speed,” and that pace is putting real pressure on Western brands. Nefkens even suggests that import restrictions could backfire for US companies by insulating them from the intense competition and rapid innovation happening in Asian markets.

With tech firms, automakers and tier-1 suppliers all chasing AI revenue, the next few years are shaping up to be a fight. Nefkens expects a tough battle over who controls the money side of AI in cars. The big question is whether Western automakers can truly shift to a software-first mindset and keep up with competitors in the East.

Read the full article at Automotive World.