An educated citizenry is a vital requisite for our survival as a free people.
- Papers of Thomas Jefferson¹
Like a musical maestro who seems to possess an inexplicable air of magic, artificial intelligence is often attributed with a magical quality beyond human comprehension.²
While this misattribution may be acceptable in the world of music, it is dangerous in the context of AI in a democracy, where fundamental decisions must be made by its citizenry. Without a familiarity with AI and how it works, how can a democratic population make informed decisions? Further, how can such a population protect itself from exploitation by ill-intentioned corporations and other entities?
Let’s start by clearing up confusion. AI is not magic; it is applied calculus, which applied destructively can be dangerous.³ In 2010, a trillion dollar Flash Crash led by auto-trading bots, an implementation of autonomous AI, caused a 998.5 point drop in the Dow Jones…in 36 minutes.⁴ In an even more insidious manner, AI assisted decisions can lead to outcomes that contravene our basic societal values, like Microsoft’s AI, Tay, which became wildly racist within days of going online.⁵ AI can reinforce existing biases, as it did in Idaho, where the ACLU is currently fighting for the rights of 4,000 Idahoans, who without justification had their Medicaid assistance cut by 20–30%.⁶
Despite its potential negative consequences, however, AI has boundless positive potential. Some people’s limited encounters with AI in the media, with its sci-fi Wizard of Oz false feeling of power, may underestimate AI’s possibilities. The same computer that ostentatiously beat the world’s best Jeopardy players in the day cracks medical mysteries at night, curing and diagnosing diseases left and right.⁷ ⁸ ASIMO, a humanoid robot powered by AI, will revolutionize search and rescue missions.⁹ Uber, Google Maps, and even commercial airline flights use variations on AI.¹⁰ ¹¹ ¹² Tesla, BMW, Ford, GM, and many others have released autonomous car programs, which brings me to my final point.¹³
There is a classic and difficult question in AI. Imagine a Tesla Model X is approaching an intersection. It’s foggy and the car’s sensors can’t see a crowd of people until it is too late. The car decides in a fraction of a second: should it drive into the crowd killing bystanders or swerve into a pole, killing the car’s occupants? Without an informed populace making democratic decisions, car companies will decide the answer.
Despite the complications, we can’t ignore this problem.
It is estimated that autonomous AI cars could save 29,447 lives per year.¹⁴ There is a balance that can only be evaluated when we understand the underpinnings of how AI works and what it can and cannot accomplish. Elon Musk understands the great possibilities AI presents but he also warns that we must regulate it “…before it’s too late.”¹⁵ On the other side of the ocean, the UK parliament has discussed many of the issues around AI but has decided to wait before taking decisive action.
An educated stance on AI is necessary for making good decisions as citizens in a democratic society. AI is here, and it is making decisions about what you see in the world, whether you are eligible for a bank loan, and how you move through the world. AI is saving lives in Massachusetts General Hospital, saving lives at the wheel, and flying thousands of travelers through the air every day. Yet, the number of people who can properly describe its consequences, behavior, or internal functioning is but a handful.¹⁶ An educated citizenry is vital for survival as a free people.
Let’s begin to really focus on educating the American people about AI.