Is AI Limiting Users' Web Site Choices?

Artificial intelligence is either going to wrest control of advertising and marketing back from the grips of the consumer — or it is going to destroy civilization. At least that’s what you might think after reading the tsunami of discourse on AI since CES 2017.

But, as with most things, if we fail to learn history, we are doomed to repeat it. AI is not the first technology going to destroy life as we know it. E.B. White called radio “a godlike presence” coming into our lives and homes. TV would turn us into zombies, and the Internet would turn us into isolationists (well, the jury is still out on that one).

I was struck by a quote I came across recently about how content  “shapes and promotes…[and] dominates their sources of information. [C]ontinued exposure to its messages is likely to reiterate, confirm, and nourish (i.e. cultivate) their values and perspectives.”

While this quote so aptly describes media consumption over the past election cycle, it was actually written in 1986, in the book “Perspectives on Media Effects." That book describes the essence of cultivation theory, which  asserts that heavy television viewers end up shifting their world view and perceptions based on what they watch on TV. TV begins to inform how they view their own social reality.



With digital media becoming the medium of choice, surpassing TV in time spent, it is fair to expect cultivation theory to apply to digital media as well. What makes this phenomenon even more powerful in the digital world is the exponential force of choice.

The main difference between then and now is that the theory was developed in a world where “television provides a relatively restricted set of choices for a virtually unrestricted variety of interests and publics. Most of its programs are by commercial necessity designed to be watched by nearly everyone in a relatively nonselective fashion,” according to "Living With Television."

Commercial necessity is very different now, when virtually anyone can create, publish and earn income from their content.  This has created an endless — some argue infinite — number of choices online.

That brings us to the paradox of choice (popularized in Barry Schwartz’s 2004 book of the same name). He cites research done by Sheena Iyengar and Mark Lepper which found that when grocery store shoppers were faced with fewer choices of jam, they were actually more likely to purchase a jar than when they had a wide array of flavors to choose from. Greater choice resulted in shoppers being paralyzed by indecision and making no choice at all.

We are beginning to see this happen in our media decisions. As the number of available television stations continues to increase (Nielsen reports the average television viewer has 205.9 options), the number of stations viewed, on average is decreasing, to less than 10% of available channels.

We see a similar pattern in Web site traffic. According to comScore, the top five sites garner almost a third of all page views in a month ,leaving the other 999,999,995+ websites to duke it out for the other 66%. 

As choice has exploded, the conventional wisdom has been that this increase in choice put the control in the hands of consumers. But has choice actually made us more passive media consumers? Rather than seeking out our variety of options and making informed choices, the reality seems to be we revert to the tried and true and reinforce our existing behaviors, and beliefs. We’re consuming more media than ever but we are consuming more of the same.Instead, the information explosion expanding our knowledge, it is, to paraphrase, reiterating, confirming, and nourishing our individual values and perspectives.

AI is here and it will not destroy us — just as radio, television and the Internet never did. Nevertheless, many fear that AI will only exacerbate user passivity by learning what kind of content we like and directing us to similar things.

Still, machine learning (aka AI) starts with human teaching. We can use the algorithms to feed us content that challenges, questions and feeds our values and perspectives. And when we do, we won’t have to worry about history repeating itself.

3 comments about "Is AI Limiting Users' Web Site Choices?".
Check to receive email when comments are posted.
  1. Craig Mcdaniel from Sweepstakes Today LLC, May 4, 2017 at 8:53 p.m.

    Beth, this is interesting.   I have a couple of questions. is "AI" a new version of "Search" or search version 2? And if this is truly "Artificial Intelligence", will a computer be ask totally indepentent questions on it's own with only a limited about of information? 

    Right now, I can search Artifical Intelligence in the search engines. The results are determined by the amount of data in the search engine. With a true AI program wouldn't it possibly have a different a different answer or just give a different interpretation of the search results?

  2. Beth Donnelly Egan from Syracuse University replied, May 5, 2017 at 11:09 a.m.

    Hi Craig - AI, as a broad term, can be applied in many different ways to many different industries. But when it comes to media and publishing, I agree that in some ways you could call it Search 2.0. The consumer doesn't necessarily actively search but if they are fed content based on indicated interests, then that is a form of search.
    Much of the fear around AI is that it will begin asking its own questions. Us older folk liken it to HAL in 2001: A Space Odyssey. Younger generations may prefer the computer in her, the Spike Jonze film. 
    However, what I am primarily referencing here is how AI can be used to fuel the pages we are fed online. In that sense, publishers are looking to feed content that they believe will be more interesting to you and keep you on the site longer. In that case, the consumer hasn't actively searched for something.

  3. Craig Mcdaniel from Sweepstakes Today LLC, May 5, 2017 at 1:52 p.m.

    Beth, When I entered sales and marketing 35 years ago, I had the opportunity to work with some very smart and tallented people. I wrote down hundreds of their questions. Mainly the style of the question itself and how a different question can get a different response. I continued to write for many years all types of questions and responses. I should probably create a book on the subject but never have.

    The point I found is AI can at best, ask basic questions but can not ever get the "Hal 2001" stage of being able to advance to the point of interrogation. In short machines can store massive amount of information but only humans can properly decend the information. 

Next story loading loading..