(This column originally appeared in Forbes)
To write this column my plan was to research how AI is impacting manufacturers in 2024. Spoiler alert: get ready to be underwhelmed.
Here’s what I found: there’s a lot of hype and a lot of rosy predictions right now. A writer at TechTarget lists “10 AI Use Cases in Manufacturing.” Another blog talks generally about “The Increasing Use of Technology in the Manufacturing Industry.” Research firm Forrester paints a picture of the “future” for manufacturing AI. A manufacturing group’s site lists the “manufacturing technologies that’s shaping the industry.”
It’s all very general and it can exciting. But it’s not now. What about now? What about the next 12 months? What’s hype and what is the reality of AI for manufacturers in 2024? Here’s what I’ve learned.
For starters, it’s a big company thing.
For all the hype about AI, it’s still a very big company thing. Sure there are surveys like this one that say that small businesses are getting benefits from AI. But this is literally scratching-the-surface kind of stuff. While small businesses are using simple Generative AI products to write blogs or create policies, big companies like Klarna are spending millions of dollars building AI chatbots that are replacing hundreds of customer service reps. BMW and Amazon plan to introduce “humanoid” robots powered by AI brains to “free up” workers and Tesla is almost halving its production time to build cars using AI-powered 3D printers. Corporate investment in AI “is surging” says the Wall Street Journal.
The list goes on. AI is a big company game right now. Eventually these applications will become commoditized and affordable for the rest of the world. But not this year, for sure.
Today’s AI’s manufacturing solutions are custom solutions.
There are lots of apps for creating profile pictures, writing essays and concocting recipes. But there are no significant “apps” for manufacturing worth mentioning right now. There are no big advances in AI functionality built into mainstream manufacturing, ERP, order entry or inventory management applications. As I wrote above, this will happen. But not yet. Corporations with big budgets are spending big with firms like Mobidev, Innover, Pleora Technologies and many other development to build their Large Language Models with integrated data that can be custom trained on the organizations’ internal processes and proprietary information.
Someday the software companies that serve most businesses will have these AI-driven tools available as part of their feature set. But if you want software that truly leverages AI in your manufacturing business today you’re going to need to build it yourself.
Manufacturing AI is leveraging existing technology.
AI isn’t creating anything new right now on the plant floor. Instead it’s being used to upgrade existing software and hardware. A quoting solution that’s been around for decades is using AI to construct better queries on its existing architecture. A robot that’s been used for sorting and palletizing items on a production line is now getting an AI upgrade to better identify and adapt to those products. An existing drone line can now do more independent things because an AI drives it. Big brands are now slapping sensors on the machinery and equipment they’re already producing to monitor vibrations, humidity, heat and a number of other factors that’s reported for maintenance or self-remediated.
The hardware and software are the same. But their brains are getting an upgrade.
GenAI has its limitations
Generative AI is certainly powerful. But it’s just getting started. We’re only in the first year or so of the LLM phase, where models are being built, tested, trained and — most importantly — litigated. GenAI is limited because it’s immature. It also carries a lot of risks ranging from hallucinations (getting it wrong) to misinformation to losing control of itself. Today’s AI applications are built around the GenAI model and — a year and a half after OpenAI’s groundbreaking release of GPT 3.x — they’re already looking old. Even OpenAI’s CEO Sam Altman recently admitted that the current version of ChatGPT kind of sucks” when compared to new releases on the horizon. Both Altman and chip-maker Nvidia’s CEO Jensen Huang have said that Artificial General Intelligence (AGI) — the holy grail of AI where the software performs as good as, if not better, than humans — is only a few short years away.
Altman and Huang like to downplay the hype around AGI but I believe that when this happens, AGI-driven machines and robots will be able to truly do what humans do and with these capabilities will significantly disrupt everything from assembly to formulations to materials movement on the shop floor.
The best place to see AI is on the road.
It’s still early days for manufacturing AI. But hardware companies, software companies and equipment makers are investing heavily. And for manufacturers the best place to find out what’s coming is to talk to your vendors…and go on the road. Visit conferences like this one, this one and this one — among the many others in the manufacturing space. Go to your industry trade show this year. Stop by the exhibitors’ booths.
In the past, you could miss one of these events every year or two. But not this year. This year companies are rolling out new products and upgraded technologies that are leveraging AI and the best way to get up to speed on them is to go and watch them in action.
That’s the AI reality of 2024 for those in the manufacturing space.
