Learn how IoB fuses AI and IoT to influence our behavior and how we interact with the world.
Does AI influence behavior?
As a producer of DIY videos on the Internet, I'm no stranger to the possibility that videos prominently featuring my hands may be labeled as "sensitive content" on Twitter. Why? It seems that a presumably well-meaning AI reads the proportion of flesh coloring in such videos as an indication of something untoward and thus feels it necessary to warn others about it.
It's a common phenomenon that's led video-makers to often feature gloved hands doing intricate work to avoid this trigger. While a tad absurd, it's a logical response to AI's misguided prodding. Just as Twitter analyzes videos, the Internet of Things (IoT) collects a wide range of real-world data to intuit what we're doing in the "real world" for inspection.
How this data is amalgamated and interacts with us has become known as the Internet of Behavior, or IoB. Using information, AI is presumably trained over time to serve us better.
But perhaps we should be asking: Are we training AI to match our desires, or is it training us to do what it thinks is best?
With the world's information at its fingertips, maybe AI does know best. Or—sans the subtle context clues and value judgments that humans seem better equipped to process—this becomes the prelude to various sci-fi works where humans struggle to overcome their machine overlords. Perhaps various AI incarnations are actually supposed to work in the best interest of the company that sets them up.
So, is data-driven AI promoting a good influence on human behavior, or is it on a misguided path? While it's perhaps impossible to answer that question definitively, we can start the thought process by looking at a few of the technologies that already make the IoB possible.
You can also check out our broader explanation of what Internet of Business is.
Technologies behind Internet of Behavior (IoB)
IMU (Inertial Measurement Unit): Senses motion to perceive everything from picking up your phone to taking a walk.
GPS: Senses your position on the earth.
Bluetooth Beacons: Pings smartphones via Bluetooth Low Energy (BLE) for things like advertising and indoor navigation.
Vision Sensor (Camera): Takes photos and video, which can be used for facial recognition and other purposes.
Microphone: Once the fundamental input element of a "phone," it's still useful for voice communication and passive listening.
What's the common denominator in this technology? These sensors are all housed in smartphones, which seem to be the main conduit for our collective IoB presence.
IoB knowledge can collect your phone data with your general web/social media/shopping history, along with input from "traditional" IoT devices (like Cams, Dash Buttons, Smart Locks, and Ring Doorbells) to create a behavioral model for you.
IoB: Affecting your actions
If IoB knowledge determines both what you are doing and what you may do, consider what the full use of IoB might encourage. Affecting your actions is the next logical step.
Consumer Behavior: From a business perspective, the most obvious use of IoB is encouraging purchases. Did you walk by a toy store around the time of your niece's birthday? Perhaps a coupon magically appears on your phone to encourage your patronage. Just leave a restaurant? Your phone requests a review, which feeds your personal IoB model.
Health Metrics: With monitoring hardware on your wrist in the form of a smartwatch and an IMU in your pocket to measure how long you've been in one position, IoB apps could prompt you to stretch intermittently. An app could even suggest you consult a doctor if it finds something amiss. Learn more about the use of sensors for personal health.
Automotive Insurance Rates: IMU and GPS data can communicate to share where, how far, and how aggressively you're driving. This data can then pinpoint insurance risks and estimate rates based on your actions. You may begin rethinking that impromptu drag race or consider whether driving at an excessive speed is worth it.
Better Financial Habits: AI that's focused on your best interests might encourage you to save more money for a rainy day or perhaps start riding your bike to gain both the physical and financial benefits. Unfortunately, this may be at odds with the need to encourage purchases. How various IoB priorities fight this one out is to be determined.
Affecting Real World Speech and Movements: Quite a fuss has been made about online censorship and the correct course of action. Imagine, however, that since AI tracks your movements and behavior, that you don't go to see a particular politician speak or text your friend what you think about something. Maybe you even stop discussing issues that matter to you in "private" because your phone is listening. Even what news outlet you passively watch could be tracked (via smart TV or microphone cues). Having an opinion about an issue on your own would therefore be subject to scrutiny.
These potential IoB-influence issues can be benign on the surface but become important on a personal level. Society-wise, they could weigh down quite heavily. The real question is, who or what is in control? What should be considered when designing systems to allow individuals to be in the driver's seat? How about when choosing what IoB applications to use?
Here are a few ideas:
1. Work specifically for an individual's self-interest. The recurring theme of IoB is that data-driven wisdom often pushes us to purchase something. This could mean, for example, that a drug maker could influence advice for maintaining healthy cholesterol levels, or that the ice cream shop you're walking by messages you to tempt you to pop in and purchase a few hundred unnecessary calories. To lessen this influence, what if an "AI secretary," so to speak, policed your apps for your best interest? It's a solution that would need a high level of trust and personalization.
Something like this couldn't be made for free, and consumers would need to be willing to pay for it with money rather than data. As a society, we aren't accustomed to spending money this way, and I'm not optimistic that it will happen. Barring that, one should at least keep this motivation in mind when developing or using an application.
2. Don't Limit People's Exposure. While settling into a particular routine may be most comfortable, random encounters with people, places, and things can bring new and exciting experiences. Managing this randomness in an AI that is, in theory, trained to maximize our comfort and well-being is something that must be considered.
3. Be Accurate, and Don't Overstep. Trusting an AI to look out for your best interest takes faith, but what if it gets things wrong? What if it possibly nudged people into an area where the data isn't fully developed enough to make a good decision? Suppose the feedback is incorrect in one area (i.e., labeling bare hands as sensitive). In that case, you may be less likely to listen when it nudges you to do something helpful, like visiting the doctor when health metrics indicate it's time to do so.
Another AI anecdote
On two occasions, I've pulled out my phone to make an appointment and noticed the time I was discussing with the office manager had already been selected. Per user agreement, Android phones have the right to record conversations that happen within earshot of the device, so this being less-than-random seems plausible.
Perhaps I should consider getting an iPhone instead. The subject of a New Yorker article touts Apple's improved handling of privacy. Unfortunately, that article is from nearly a year ago, and in the name of safety, Apple is at least considering scanning images for offensive content.
Given the new image-scanning proposals, perhaps we'll see a future trend where people wear blue rubber gloves in family photos. Or, maybe humans will accept that it's convenient and beneficial for a phone to select the correct appointment time without prompting. In turn, AI may accept bare hands in public as socially acceptable.
The future of IoB software hinges on developers who will consider the needs of individuals. But it also depends on consumers: We must understand the need to prioritize apps that work for both our collective and individual good.