Let’s hope that Bill Gates latest prediction, that “in the next five to 10 years, AI-driven software will finally deliver on the promise of revolutionizing the way people teach and learn”, is more accurate that when, in 2006 he said of traditional advertising, “I’m stunned how people aren’t seeing that in 5 years we will laugh at what we had”. Addressable advertising didn’t kill mass marketing, but we should apply some lessons learned from our experience with it to AI, or AI might kill us!
AdTechnicians hooked up a flashy new advertising apparatus and the most expensive advertising experiment in history was underway; serving the “right ad to the right person at the right time”. The format to support that notion relied mainly on banner ads, which, during a good campaign would achieve a whopping 1 click in 10,000 (CTR has since been retired fired). Not only were consumers not paying attention but were actually getting annoyed. Pop Up’s were a distraction and general nuisance to consumers who were trying to engage with content on websites which ultimately lead to them abandoning the publisher all together. This was the Dark Matter of online advertising and although no one has ever measured its true impact it’s safe to say that what once claimed to end the interruption model of advertising took “annoying and intrusive” to stratospheric new heights. More importantly, 1 to 1 advertising is ‘Private’, while 1 to many is ‘Public’ and public advertising does 3 things for Brands better than private advertising can – make them legitimate, speak to their substance and create a cultural imprint. The expensive act of mass marketing lets people know a company or service means business. Mass marketers can’t afford to be crummy so consumers feel a high degree of risk aversion. Products only have certain qualities or personalities if they exist in the collective psyche. All of that only works if large populations are confident they are taking part in a shared experience.
But the conversation was all about the tech; the inputs and outputs, DSP’s, SSP’s and ad exchanges. We marveled at root systems, nested logic regression, algorithms, data taxonomy, dynamic ad insertion, real time bidding, dashboards and efficient workflows. Something was missing, however. Along with fraud, survalence and the ill effects of social media on society that came with internet advertising – it turned out not to be very dependable, secure, trustworthy, or aligned with human values. We were so blown away by the technology that we overlooked the humanity.
You best believe we ought not to make that same mistake with AI.
Digvijay Wadekar of the Institute for Advanced Study in Princeton said of AI, “…machine-learning focuses on deep neural networks. These are very powerful, but the drawback is that they are almost like a black box. We cannot understand what goes on in them.” We’re currently in the stage with AI where the common thread in most of the News generated about it is concerned with how it makes us feel. We’re excited… and anxious. We’re hopeful…but terrorized. Mostly this is because of the unknown. We didn’t understand the black box of programmatic advertising but were nonetheless rather cavalier about it. That can’t happen with AI. A study led by a professor from the University of Central Florida has identified six challenges that must be overcome in order to improve our relationship with artificial intelligence and guarantee its ethical and fair utilization, the tenets of which require it be dependable, secure, trustworthy, and aligned with human values. Sound familiar?