Commentary

Don't Underestimate AI Just Because It's Overhyped

This AI Insider column from early 2017 gave MediaPost readers the inside scoop on how Cambridge Analytica built up its data and algorithms to exploit Facebook’s massive user base — a year or more ahead of the mainstream media. Note the last line about "the big-brother-like political impact of big data": "It is frightening."

To paraphrase Joseph Heller, just because AI is overhyped doesn’t mean it won’t transform the world — faster than you can handle.

I remember sitting in a conference audience in the late 1990s during the fat part of the first dot-com expansion curve, when everyone was complaining that the Internet was irrationally overhyped. Then, pre-Google Eric Schmidt took the stage and told us that, “I actually think the Internet is underhyped.” As a tech journalist in those days, I’d had the privilege of long talks with Schmidt and hadn’t wasted the opportunity to learn. Other people laughed, but I knew he was serious — and he was right.

The point is, I’ve begun to get the sense that most marketers aren’t yet taking AI seriously enough. Sure, it won’t put you out of business next month. It won’t make everything you know wrong, overnight. But consider the implications of Bill Gates’ aphorism from "The Road Ahead”: “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next 10.” No, really, stop -- and think.

Ten years is not so long a time. Not when you’re talking about building mountains of clean, usable data (the fuel of AI); or disrupting your still-new digital marketing approaches; or transforming customer experiences; or redefining job roles; or restructuring your org chart; or reinventing your own career arc. Again.

What started me down this thought path was a report on AI released Tuesday by Altimeter, and a story about how big data and algorithms probably decided the 2016 presidential election. Both are compelling, must-reads.  

Altimeter’s  “Age of AI: How Artificial Intelligence Is Transforming Organizations,” written by Susan Etlinger, cuts through the hype in simple prose that lays out AI’s history of fits and starts, explains why this time is different, and explores current real use cases.

So why is this time different? Here’s what Etlinger wrote: “Given the number of false starts in the past several decades, it’s reasonable to wonder why now is different. There are three key factors that distinguish today’s AI climate from that of the past:
• Massive and available datasets (also known as “Big Data”);
• Inexpensive parallel computation; and
• Improved algorithms.
The combination of these three factors has made it possible, finally, for AI to become not just a wild idea or rarefied technology, but a commercial reality.”

She goes on to explore uses cases and potential pitfalls, such as unintended consequences, hidden biases in datasets, and ethical dilemmas. We agree about how critical AI/big data ethics will be in the near future, but that’s not her focus, so my related post  provides a bit of a deeper dive.

She also notes that: “Products and services will, by their nature, require public use (and in some cases, such as the incident with Microsoft’s Tay, abuse) to learn how to navigate unforeseen circumstances. For this reason, organizations contemplating AI-powered offerings must balance the imperative to ship product and learn from it with the imperative to preserve and protect customer experience and brand.”

That’s a clear exhortation to get started, ASAP, people. With AI, you have no choice: you cannot get ready, aim and fire. You have to fire in order to aim.

Then she points out that it’s important when training AIs to use factual data rather than opinion, mainly because you introduce less bias that way. For marketers, in particular, “transactional and behavioral data should trump demographic data.”

That one word, “demographic,” spun me back to “The Data That Turned the World Upside Down,”  the Motherboard story about how big data and algorithms probably decided the 2016 presidential election. While Trump and Clinton both had huge social media datasets, and strategies for big data and social media, this story dives deep into Trump’s use of psychographics as opposed to demographics. And, oddly (or not so?), all managed by the same company that helped market Brexit in the UK: Cambridge Analytica.

The psychographic models used in both cases assessed people on five personality traits: openness, conscientiousness, extroversion, agreeableness, and neuroticism. The story explains, thoroughly but simply, how the original researcher built his psychographic models over several years, working with ever-growing datasets provided by Facebook users, matching those traits up with profile data, until the realization of what was possible kind of freaked him out: “[n]ot only can psychological profiles be created from your data, but your data can also be used the other way round to search for specific profiles: all anxious fathers, all angry introverts, for example—or maybe even all undecided Democrats? Essentially, what Kosinski had invented was sort of a people search engine.”

It then goes on to explain in equally thorough fashion how these capabilities were applied by the Trump campaign.

So that was a hard transition -- from exhorting you to take up AI in your marketing organization, ASAP, to the big-brother-like political impact of big data. But no, not really. “The Data That Turned the World Upside Down” is simply the case study that shows exactly what big data already can do, today, in the hands of the right algorithm. It is frightening.

Next story loading loading..