Think More, Click Less to Thwart AI Manipulation – March 2022
In Christopher Thompson’s recent (Union Leader) column -- “You can choose whom you do business with” -- he points out that private businesses like Go-Fund-Me can establish their own guidelines, in this case excluding support for protesting truckers. He didn’t address the challenge of cancel culture pointing out the Citizens United ruling that gives private corporations freedom of speech, censorship and potential advocacy towards political objectives. He also did not point out the pathway for abuse that social media provides for entities that do not have America’s best interests at heart, who stoke the fires of dissent and division.
The trucker’s protest might not have been started by foreign entities from Russia (the IRA), China, ISIS, or others, but there is little doubt they have contributed to, and amplified, messages on all sides of this, or other issues. They'll do anything they might discredit democracy and undermine our economy.
Social media companies generate massive profits selling access to us to the highest bidders. For more details and an example of this, I recommend the Netflix documentary “The Social Dilemma”. We are the product being sold, and this is done with personalized portfolios on each of us built on massive online data collection and analysis by artificial intelligence (AI) applications. It should not be a surprise to find these companies have characterized each of us in terms of gender, income level, age, marital status, political affiliation, location, purchasing preferences or any other conclusions one might draw from our search patterns, online profiles, credit card purchases, property records, car registration records, or other public/purchasable data.
You might be surprised to find, based on research of Stanford psychologist Michael Kosinski, that your online activities also reveal your psychological profile, gender preference, and other insight -- including things you might not know about yourself. All of this goes into the AI black-boxes that target us for a particular advertisement, headline, or message. This focused amplification of messages is protected by Section 230, which shields social media platforms from liability.
You might say “I don’t use Facebook” but that is half true. While you don’t use Facebook, Facebook uses you. “Free services” collect information and target advertisements across the Internet. When you see a “Like” button, or “share on Facebook” (or other platforms) the presence of these logos on a site are a visible indication of their tracking.
There are also invisible ways to track you, so recognize that you are being tracked. While you can initiate a purge of data you have shared with specific sites, you can’t purge the data they have contractually acquired, nor the profile they have developed using AI algorithms about you as an individual to target. This data is their corporate “crown jewels” that they do not sell. What they do sell is access to you using your online profile. When a car company wants to sell a car, they know just how to reach a target community and the spin that will trigger action in that community.
Nobel Laurent psychologist Daniel Kahneman, in his book “Thinking Fast and Slow” demonstrates that even if you think you ignored ads, they registered and influence your thinking. Repetition builds acceptance of even false information. And we can be “primed” for decisions in subtle ways that I pretend don’t work on me.
Social media companies do not need to read the research, although some employees have. Some employees have blown whistles, left and started resistance movements, or even created documentaries and TED videos. Entities like the IRA in Russia, and similar organizations in China, Iran and even the U.S. have studied this and identified how it can advance their objectives. When I suggest disruption campaigns have foreign advocates, I am suggesting armies of well-funded, well-informed experts who have found ways to use our freedom of speech against us.
This begs the question of “is the best protection against free-speech abuse, more speech?” This is less clear than I would like. Lies travel “farther, faster and deeper” than the truth as demonstrated by MIT’s Sinan Aral. This makes counter-programming difficult. Kahneman’s work suggests one key is to start with disbelief, try to disrupt our intuitive knee-jerk reactions and engage in the arduous task of evaluating information before we act. Since fear and outrage are the strongest attention grabbers, even better than kittens, and these appeal to our basic instincts, this "take your time and think" process is non-trivial.
But here is the real problem. The AI capabilities used to beat humans at chess, are now being applied to beat humans at the game of free-choice. The black-box selection of individuals to target and the real time adjustment to capture your attention and adjust the message to affect your thinking and action are developing beyond human understanding. For now, our only protection is to “think more – click less.”