3M: ‘We only use AI for complex problems we can't solve any other way’
Paul Cardno, 3M’s global digital innovation lead on rolling out AI responsibly
“This technology is really early-stage stuff. If there weren't billions of dollars of investment behind it a lot of it would still be living in labs," said Paul Cardno, global digital innovation leader at industrial giant 3M.
"We're at the stage where that scale-up work goes on.”
But, he added, GenAI is not conforming to the usual adoption curve: "Normally technology gets used by 10% of people at this stage, whereas adoption is already in the 30s."
This combination of immaturity and widespread use creates a governance headache around shadow AI, copyright, bias, transparency, cross-jurisdictional compliance and ethical issues.
Because of this, and in view of its heavy resource consumption, GenAI is deployed within 3M only where nothing else will do the job. "We only use AI for complex problems, when we can't solve the problem any other way. If we can solve complex things in other ways we will."
Cardno was speaking during a debate in London last week organised by 3M and The British Science Association, hosted by accelerator Digital Catapult about the UK's role in AI innovation. He said he was keen to see the UK take a lead in providing investment incentives and regulation around the use of AI, particularly around sustainability, where any potential benefits are currently at odds with its energy and water requirements.
"There are clear benefits around green jobs, around making sure that we can adapt to the challenges of climate change.
“But with any new technology that's emerging it's always quite consumptive. The key thing is making sure that there's a framework that puts in place the incentives to make it less consumptive, to make it smaller, to make it more effective and more efficient."
‘Business likes certainty’
The UK AI Bill has yet to see the light of day. Panellists speculated that the government may be waiting to see what the incoming US administration does before it decides, but this is becoming frustrating to businesses.
"Business likes certainty," said Cardno. "We like to know where we're investing, and good regulatory frameworks can support that good governance that allows us to be faster and more innovative because it gives us the space we need.
"In the UK we're still waiting for that regulation to come out, but we've seen real progress in the US, China and the EU."
That said, as an early adopter, 3M has created its own governance framework for responsible AI. "There's always a space for organisations in lieu of regulation to make sure they are being responsible, because at the end of the day it's about managing risk."
To this end, 3M has built its own internal AI platform called 3M Navigator, based on Microsoft technology and OpenAI. This is made available to all employees and can be monitored, with guardrails and guidelines put in place.
"So, it doesn't always involve regulation," said Cardno. "But [regulation] certainly helps us there and makes things clear for us."
‘People that use it are so much happier’
As for use cases, Cardno mentioned optimising complex manufacturing supply chains as one of the complex areas where AI is starting to make a real difference.
But it’s not just data-heavy industrial tasks that can benefit. Indeed, Navigator is already helping with jobs that require subtlety and nuance, he said. GenAI's rewrite of product content on 3M's website has resulted in a 25% rise in engagement.
As another example, when a Japanese team member writes emails to colleagues in the US, she asks the tool to reformulate them in a way "that will culturally land better. That's really made a difference in some personal productivity spaces."
But perhaps the biggest benefit has been in employee satisfaction.
"My favourite bit is that the people that use it are so much happier. They're saying, 'I'm doing ten times the work I used to do, the results are better, and I'm enjoying it more because I'm not having to do the kind of nuts-and-bolts stuff I was doing before’.
However, this has not happened overnight, Cardno emphasised.
"There's been a lot of work to get to this stage, to generalise the system, but once you go beyond that you can start to transform how you work, and how you engage with customers."