The biggest mistake made regarding AI is thinking it’s automatic, autonomous, and doesn’t require humans to succeed. And even more to the point, when it comes to programmatic, that we’re engaged in a battle between AI and human qualities – some sort of Terminator-style fight for supremacy.
While I would be the last person to recommend hand-to-hand combat with a T-1000, I do have a fresh view on the old ‘man versus machine’ debate which I believe can really help us all progress.
Beware The Self-Driving Ad Campaign
AI-driven media buying has of course allowed brands to do a whole lot more with less. And whichever way you look at it, programmatic has amplified choice around inventory, and campaign targeting.
But AI resolutely is not a Tesla ‘self-drive’ mode for business. In fact, it is this misunderstanding that is leading to all sorts of problems, both in advertising – and increasingly in society as a whole.
The Facebook/Cambridge Analytica fiasco is the most prominent recent example. The social media giant supplanted most publishers as a news destination, but refused to accept the normal responsibilities of a publisher, dispensing with human editors. So, as a ‘platform’ it could rely solely on AI to separate lies and propaganda from valid information. Not to mention monitoring its ad platform for bad actors. I don’t think I need to add how well this all played out.
From Filter Bubble Back To Shared Experience
Another notable learning from Facebook’s ills is how AI took the company so far from its stated mission – ‘bringing the world closer together.’ But if anything, the Newsfeed algorithm, primed to reward engagement without any moral compass, pushed people further apart, into their own likeminded micro-ghettos.
Taking a step back, you can see the same thing happening more generally in programmatic display. In place of the news algorithm, we have a slavish reliance on micro-targeting, and third party data segments built on historical signals of questionable authority. The net result: consumer animosity at being over-targeted or retargeted uncontrollably – essentially stuck in their own ad wormhole.
Lest we forget, advertising is a human business. Human intelligence, cultural understanding, humour and shared experiences win the day. This all seems so obvious, yet all the evidence suggests it’s even easier to overlook.
Because what we see far too often in digital campaigns are those human considerations being supplanted by machine-like qualities. In other words, that someone has put media buying into ‘self-driving’ mode – and the consumer’s only hope of escape is to reach for the ad blocker.
Getting AI right was central to the report delivered by the recently created Government AI Committee. The general conclusion was that AI must be for “the common good and benefit of humanity” as well as warning against allowing AI systems to “hurt, destroy or deceive people.” In our world, the challenges around deception are deepening, and the need for clarity on why and how you are speaking to consumers is increasingly important. Indeed, in a post-GDPR world, we need to be clearer than ever on how we manage customer data, not least where it connects to machine learning and AI.
Within these boundaries, we can and still should keep evolving our usage of AI to understand what our customers do. But as marketeers, we should also apply empathy to really understand why they do it too.
For processing power of course, AI beats human every time. And a whole range of other repetitive tasks can already be successfully automated to facilitate better media buying. But empathy isn’t one of them. In the words of the Director of MIT’s AI Laboratory Daniela Rus:
“Artificial intelligence comes nowhere close to people in terms of perception, reasoning, communication and creativity.”
I believe the future is in blending the power of smart technology, with the many human qualities an algorithm can’t provide. Effective media buying that also promotes a positive customer experience is within our grasp.
Jacqui is speaking at our upcoming Digital Marketing Summit on 26 April 2018. For more information regarding the event and for the full agenda please see here.