So What Drives The Cars Then, Monkeys?

UK consumers are somewhat confused by AI. Which is not a surprise because the technology is moving fast. But the speed of change is not so much the problem. It's more a matter of basic comprehension about what AI is and does.

More than two thirds of UK consumers agree that “any public content that’s created with AI should be labeled as such.” They know AI exists, they are somewhat aware of its emerging capabilities to generate new content, and they what to know when they are talking to a bot or looking at machine-generated content.

Yet when we ask consumers where they think AI is currently deployed, a majority conflate chatbots with AI, while only around a third consider that product recommendation engines might be using machine learning to spot patterns in their browsing behaviour to suggest new things for them to buy.

Perhaps more bafflingly though, less than half think that self-driving cars are powered by AI – a fact that promoted my colleague Audrey Chee-Read to ask the question I stole for the title of this post, “so who do they think is driving the cars then, monkeys?”

Joking aside, what this pokes at is a tension which business leaders must confront when building AI powered customer experiences. First, your average consumer isn’t going to differentiate between your nana's good old fashioned predictive AI and fancy-shmansy new generative AI (why would they?). Second, they increasingly expect you to be transparent about where you are deploying AI. And third, they may well act all surprised when they find out how much AI is already out there in the wild.

Ask not can we, but should we?

When we ask business and technology leaders what’s stopping them going faster with AI initiatives, ethical and privacy concerns and a lack of trust in AI systems rank as high as concerns about how fast the technology is evolving.

As of December in the UK, less than a third of our survey respondents told us they’d used generative AI. For those that hadn’t, “I haven’t had a reason to” topped the reasons why not, but lack of knowledge about how it worked tied with concerns about ethics and lack of trust in the output as the next most pressing barriers to adoption.

The key question businesses need to be asking in the short term is not can we build a customer-facing generative AI widget chat-thingy, but should we? Legally, can we, either under current, or future proposed legislation? And remember, existing legislation like GDPR already covers many aspects of customer-facing AI apps. And if the answer to that is yes, then should we given the risks of hallucination, error, bias, poor explainability, privacy and copyright breach. That’s a lot.

Clearly there’s a lot more research to be done to map consumer trust and AI. Watch this space...

TL;DR: Proceed with caution, but absolutely proceed or you’ll be left behind fast!