Will AI Make Our Kids More Human? (or Steal Their Humanity?)
July 15, 2019
If we don’t plan our AI future, we risk handing our kids a world where machines make decisions for them and bias influences everything from routine policing to insurance decisions. To prepare this generation for what is coming, we must equip them with the ability to understand, question and manage AI in the world around them.
It’s a child’s right (and a parent’s) to know whether daily interactions are with a human, an AI, or both. As a society we want to make deliberate choices about what we automate and what we don’t. We also want to think carefully about how humans and AI best work together.
Today’s eight-year olds will graduate from high school in 2030. By then there will be over 100 billion connected devices in the world. Many of these devices will be accessible to kids via voice and gesture controls powered by AI. These kids will ask their mirror for the daily weather forecast, take a self-driving car to a friend’s house and pay their robot lunch delivery service with a hand gesture.
That brings us to the critical question: “If an AI driven future is inevitable, how do we get the version we want?”
The past decade brought us smart phones, tablet computers, self-driving cars, chat bots, VR headsets, phone-based GPS, CRISPR (gene editing), 3d printed organs and drone delivery services. We’ve also seen the rise of social media, online networks and a huge variety of services that depend on combinations of these advances.
If we’ve learned anything in the past decade or so, it’s that the rate of change is increasing at an accelerating rate. This is a problem for humans. We’re pretty good at thinking about steady change and pretty bad at planning for changes that happen quickly and at massive scale.
Our struggle to understand huge changes is well captured by physicist Al Bartlett’s famous quote, “The greatest shortcoming of the human race is the inability to understand the exponential function.” The early part an exponential curve is flat and seems innocuous but then the curve ramps up and suddenly massive change is occurring over very short time frames.
The power of exponential functions can be explained via a classic story. A generous king offered the inventor of chess a reward of his own choosing. The inventor made what seemed to be a modest request: just one grain of rice on the first square of the chess board, two grains on the second, four on the third, eight on the fourth, and so on, with the number of rice grains doubling on each successive square. The king agreed the request was modest, but he was wrong. When his courtiers tried to complete the task, they found that it would require more rice than existed in the entire kingdom, indeed in the entire world!
Consider our current problems with data and privacy on social media. Almost everyone did a poor job of predicting what the rise of Facebook and other ubiquitous social media services would lead to. It’s only now we belatedly realize social media platforms could subvert democracy, harm the economy and exacerbate ethnic, class and national divisions.
AI is the next thing that will rock our world and it will make problems with social media look like the warm breeze that precedes a hurricane. Experts like Kai-Fu Lee and Elon Musk agree AI will be hugely disruptive, but they don’t agree on what form the disruption will take.
Further complicating our automated future is the fact that AI is becoming both ubiquitous and invisible. Using Siri and Alexa are already second nature for anyone under the age of 18, but few think about how these services work and even fewer think about how they should work. And that is bad.
Last year a woman in Portland, Oregon had her private conversations secretly recorded by the voice controlled Amazon Alexa in her home and then those conversations were sent to a random contact in Seattle. Shouldn’t we be loudly demanding privacy and other safeguards from the manufacturers of these connected products? Shouldn’t everyone know how they work and what they are risking by using them?
There are SOOOO many questions we should be asking…
- What data is being collected?
- How is it being used?
- Is it really necessary for the service being provided?
Most importantly, we should be training our children to ask them as well.
So, what can we do to “future proof” our kids? What skills, knowledge and strategies will give them a leg up in a world overrun with software, robots, connected devices and self-driving vehicles? And what type of relationship do we want our kids to have with AI?
I’d like to propose a simple model for helping all children prepare for our AI driven future. Starting in kindergarten, I propose all kids should:
- Be aware of how AI works and where it lives (five big ideas).
- Gain a deep understanding of what humans do best and what AI does best.
- Debate what should be automated and what shouldn’t.
Let’s dive a little deeper….
#1: AI Awareness
Experts have settled on five big ideas that all kids should know about AI:
In rough order of importance, AI:
- Is impacting society in a meaningful way already and that impact will only increase.
- Allows computers to understand (perceive) their environment, including the ability to recognize human emotions.
- Uses “agents” (programs that learn from their environment) that maintain models of the world and use these models to make decisions.
- Can learn from data and that data can be collected from an incredible array of sources.
- Will interact with humans in an increasing natural way over time.
With this knowledge kids will be more empowered to ask “why” an AI is doing something as well as “what” data is being used to do it. They can also more effectively recognize when they are interacting with AI and when AI is influencing their decisions or even making decisions for them.
#2: Jobs for People
In a world where AI is getting smarter and more efficient at an exponential rate, kids must learn to constantly sharpen their problem-solving skills and update their understanding of how the tech around them works.
This chart shows how the young and certain minority groups will be disproportionately affected by AI job displacement in the near future.
Average Automation Potential by Age or Ethnicity – 2017
But the bigger point is that even for groups that are “safer” from automation, 40% of today’s jobs are likely to go away! This is more displacement and disruption than ever seen before in human history.
So what can our kids study to give them immunity from automation? Some, like Kai Fu Lee, argue that the future will belong to creative combinations of AI and humans. For example, imagine medical AI that examines you, diagnoses any important issues and shares those issues with both a human doctor and the patient.
The doctor, free from the heavy burden of diagnosis and treatment research, can focus her energy on patient buy-in and understanding, the execution of treatment and thoughtful follow-up. In order to thrive alongside it, people need to think carefully about skills that only humans have, like empathy.
#3: What Should Be Automated?
In general humans are on a path where everything that can be automated, will be automated. Barista, cook, delivery driver, welder, legal discovery, x-ray analysis, accounting…. You get the idea.
I believe we should be actively thinking about what we DON’T want to automate. Gardening for example. Many will probably want to automate basic yard maintenance however, there will be almost as many who will want to do it themselves because they see it as relaxing and productive.
I’m confident very few of us want to live a Wall-e world where all work has been automated and humans are only left with leisure activity. Billions of people love their work and get a tremendous amount of satisfaction out of doing it well. Shouldn’t we be saving some of our favorite tasks from the robots?
An encouraging trend is recent research on collaboration between AI and humans. In a study that showed how AI’s and humans could effectively collaborate to play video games against all human teams, the authors noted “most real-world endeavors require teamwork”. And their results show that the AI + human pairing can be quite powerful and effective. It also drives home that point that AI learns what we help it learn, so let’s focus on collaboration and socially positive use cases.
This is where the optimist in me steps in…
Imagine a world where products and services are designed with maximum human happiness and satisfaction in mind. Where mindless and unhealthy tasks have largely been eliminated and the workforce is focused on wellness, entertainment, art, exploration and education.
This is a world that I would be proud to hand over to my three kids. But we’ll only get there if we start preparing now. Let’s start asking more questions, demanding more thoughtful plans and debating where we want AI to help us go.
Most importantly, let’s raise a generation that understands how software works and will think carefully about training and collaborating with AI to be sure that AI continues to be a powerful tool for advancing human achievement and wellness.