AI Can't Replace Industrial Design — Here's Why Jony Ive's Delay Proves It
Jony Ive just proved AI can't do this part of the job.
Even with the most advanced AI technology available, the collaboration between OpenAI and Ive's io has not been able to overcome delays in launching its first hardware product. Not because the team lacks ideas — presumably, they're still searching for the one that's actually right, and no amount of AI seems able to make that search go faster.
It's a useful case study for any business trying to understand where AI genuinely accelerates product development, and where it doesn't. As an industrial design consultancy in Sydney, this is a question we sit inside of constantly — and the answer comes down to a simple but often misunderstood spectrum.
Predictability and Innovation Sit at Opposite Ends of a Spectrum
The more predictable the solution, the less innovative the result. The more innovative a solution is, the more unpredictable the result. Big innovations are inherently fuzzy, whereas incremental improvements in technology — bigger, faster, lighter — are more concrete.
This distinction matters because AI performs very differently depending on which end of that spectrum a project sits on.
Engineering Sits at the Predictable End
Use AI to optimise a bracket for weight and load tolerance, and it will out-iterate any engineer in the room. Generative design tools already run thousands of variants in aerospace and automotive, converging quickly on a mathematically correct answer. This works because the target is fully specified — you know what "better" looks like before you even start.
Industrial Design Lives at the Opposite End
Nobody yet knows what "better" looks like for a genuinely new product category — a screenless AI device with no obvious precedent, for example. There's no spec to optimise against, only a feeling for what's right once someone finally lands on it. That search runs on judgment under genuine uncertainty, and it's a kind of risk AI has never had to carry.
I had a colleague once look at an early foam model I was working on and say, "I just can't see where the innovation is in this." He wasn't wrong to say it — it didn't exist yet. Ive's team is living that same discomfort right now, at a scale most businesses will never operate at.
Where This Leaves Product-Led Businesses
The real skill for any business isn't choosing between AI and human judgment — it's knowing which end of the continuum a given project sits on, and resourcing accordingly. Engineering teams and AI tools for the predictable work: thinner, lighter, faster. Industrial designers for the unpredictable work: the ideas that don't exist yet, and won't, until someone with the right instincts is given room to find them.
This is precisely the kind of work a specialist product design consultancy in Sydney exists to do — sitting in the fuzzy, high-uncertainty end of development where AI can't yet operate, and where the depth of human engagement in the process is what actually produces category-defining outcomes.
Looking for an industrial designer in Sydney who can navigate that uncertainty with you? Get in touch with Paranormal Design to discuss your next product development project.