We are not ready for AI
Rand McNally maps in your car, the corded landline phone on your kitchen wall, dial up modems, brick sized car phones, 36 exposure camera film. If you think about it, these were all ubiquitous less than two decades ago. Today you would be hard-pressed to find a single one of these.
It was the futurist Roy Amara who once said, “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
It is my belief that machine learning will fundamentally transform so many aspects of our lives, that two decades from now, our current tools and behaviors would be as obsolete as the examples above.
There have been many articles talking about the winters of past AI hype cycles, and how maybe we are now entering the trough of disillusionment of the current AI hype cycle. While we may be entering a trough, this cycle will be different from cycles of the past. I am convinced society will be blindsided by the speed of change that will soon descend upon us.
To make my point, let me paint a picture of our future for you.
In the future cheap and ubiquitous digital video cameras will capture video everywhere: on streets, inside shopping malls, inside your grocery stores, maybe you been outside your house. This by the way, is already happening. I saw this in two touristy areas, one in China and one in Zanzibar. Every street intersection had 8 cameras, looking up and down each direction. London is currently blanketed by an estimated half a million CCTVs.
Facial recognition will also be ubiquitous. People have shown over and over again that convenience trumps privacy, and establishing identity via facial verification is super convenient. Facial recognition will probably replace ID cards at companies, and maybe even at airports. Your face is your identity, and will also drive retail payments. Imagine paying your bill at restaurants and stores by simply looking at the camera.
Facial recognition can and will be used by governments as well. China is already leading the way in this. There have been several articles talking about the giant surveillance network run by the Chinese government in the Xinjiang Uyghur Autonomous Region of China.
In this scenario, there exist several companies like our fictional company EOS (the Eye of Sauron), that sell the combination of location based data combined with user identity to companies, so that you can be convinced to buy their goods at the right time and the right place. As the biggest one of these, EOS is the Facebook or Google of the future.
Instead of tracking your digital footprint across different websites, EOS tracks your real presence across the physical world aggregating it from multiple retail and public sources, analyzing and extrapolating your location history, so that it can accurately guess where you have been at any given point in time in the past. From this they can observe countless patterns in any person’s behavior, predicting with a high degree of accuracy where you would be at any given point in time in the future based on your routine. Now let us look at some increasingly dystopian examples of this tech:
By tracking you in the grocery store, EOS knows that you consume roughly 2 gallons of milk every week. You haven’t had milk delivered nor have you gone to the grocery store in 8 days. Chances are you will go to your grocery store today evening, or will need milk delivered to you.
EOS is a platform company. Other companies build their applications on top of the EOS platform. Every networking event or conference that you go to, everyone knows who you are. It is convenient and efficient. As you look through your phone camera at the room, everyone’s name title & company pops up. You select someone that looks interesting, and the app now pulls up their social media profiles and pictures, and suggests common interests that can help you break the ice.
Every sales person is now equipped with a similar app. Sales techniques now include “accidentally” running into prospects while they are out and about their daily lives. Even low profile venture capitalists and executives run into desperate founders/salespeople while out with their family for dinner, with alarming regularity.
Stalking people is now effortless. Susan runs into Joe from her yoga class in so many places that she suspects it’s no longer a coincidence.
Kidnapping has now become easier by an order of magnitude. No longer do you have to plan ahead, identify a target, and then track them and their daily behavior. Everyone’s identity and status i.e. their value is available realtime. Simply walk down the street and pick up a high-value target when the opportunity presents itself.
Divorce lawyers have their own ways of analyzing this data. Some of them drum up business proactively. Bill gets a call from a law firm: “Hey did you know that Janet has been spending time at this apartment two or three times a week in the afternoon when you are at work?”. Elizabeth gets a call too: “Did you know that Jack went to a place of disrepute three times in the last 8 months?”.
Law enforcement now has access to everyone’s past location history. For every new low-level drug dealer that is identified, police can now go back and identify all his customers. Insurance companies can now identify drivers who do rolling stops.
By now I’m sure, your mind will start to generate many more possible uses and abuses for realtime pervasive facial recognition. The point of this article is not to scare you, but merely point out some possible trajectories of our AI enabled future. The future will unfold in ways which will be extremely hard to predict, but responsible use of AI will be a precarious tightrope to walk.
Balancing customer demands for convenience, and company demands for more data to better serve customers, versus the myriad potential abuses of privacy will require self-policing and restraint from companies, and thoughtful policies from lawmakers. History has shown that neither of these unfortunately, are a given.
If you liked what you read, please feel free to share the article.