MARK HALPERIN: Social media damaged a generation. Are we about to let AI do the same?
Imagine: two hundred aliens land on Earth, with thousands and thousands more of them expected in looming, unstoppable waves.
And imagine, too, that we know some of the aliens are here to help us, to make our planet better, smarter, and more efficient. But we also know that some of the aliens are here to do us harm, perhaps injure us irreparably. What we don’t know is this: which aliens will be good to us, and which will be malevolent?
That is the situation with AI.
But it is more complicated than that.
Because the good aliens and the bad aliens aren’t necessarily different aliens. The same technology that can help discover a cure for cancer, give a low-income kid access to a world-class tutor, make government more efficient, and produce extraordinary new wealth also can deceive, manipulate, displace, addict, and destroy. The alien who comes bearing gifts in the morning might be the one who burns down the village at night.
And yet for the multitudes of op-eds and think pieces, the anxious parent text chains and furtive bot-synthesized office memos, the accelerating quotidian reliance on AI for driving directions, recipes, fun facts, and how-tos, we aren’t discussing AI with the frequency and intensity it warrants.
We are merely watching, transfixed and paralyzed; or turning away and scurrying back to landlines and paperback books; or playing along, sinking into the lazy comfort of an easy answer. An algorithmic echo of Ray Bradbury’s 1948 short story, "Mars is Heaven."
REP. TED LIEU: AI IS ALREADY TOO POWERFUL. WE NEED A KILL SWITCH BEFORE DISASTER STRIKES
There is little doubt that AI will be the next leap forward in human advancement, akin to the printing press, the Industrial Revolution, penicillin, and the internet. But we are not having the urgent, necessary conversations about what it all means.
China, meanwhile, is all in on AI, investigating its most sinister uses, learning its gifts and powers, leveraging it, weaponizing it. In the United States, however, the topic of AI veers shambolically from flippant Matrix memes to end-of-days-style apocalyptic panic, to tepid comparisons with old-school Google.
WHEN AI TESTS CAUSE DAMAGE, WE NEED STRONGER SAFEGUARDS AND REAL ACCOUNTABILITY
The implications of AI are so vast, so incomprehensible, that it is unclear which entities are equipped to wrangle it, or at least enforce some regulations, whether it is the White House or Congress, the two branches united, or state-by-state legislation. Additionally, AI is changing so quickly, and government moves so slowly, that by the time any measures are passed, the technology that prompted the hearings might be several generations old.
One can look at the rise of social media to see a clear, cautionary analogy.
What began as an innocuous way to share life updates, family photos, and dog videos grew enormous and toxic, changing the way people, especially children and teens, process information. Changing their hobbies and passions and brain chemistry. Changing the way they make friends, view themselves, and interact with the world. Social media has transformed the way we get news and connect with people, but belatedly, has induced a scramble to scale back the more malignant elements of its influence, especially on the young.
The great lesson of social media isn’t simply that we failed to regulate it. It is that we waited to see what would happen.
SEN. BERNIE SANDERS: YES, WE MUST HIT THE PAUSE BUTTON ON AI BEFORE IT IS TOO LATE
We let hundreds of millions of people use it, let the companies optimize it, let children become consumed by it, and then, years later, began accumulating the evidence of what it had done. Only after the patient got sick did we start looking for the right medicine.
Meta recently took a significant step and acknowledged its contribution to the crisis, following a massive nationwide lawsuit accusing the company of deliberately designing Facebook and Instagram to be addictive to young people, along with charges that the company was intrusive and unethical with its data collection.
UK TO BAN TIKTOK, YOUTUBE, OTHER SOCIAL MEDIA APPS FOR CHILDREN UNDER 16, STARMER SAYS
In addition to paying out as much as $18 billion in damages, Meta promised to make major changes to how children and teenagers may access and use their platforms. Meta further called upon its industry competitors, YouTube and TikTok, to enact similar alterations.
For years, millions of parents, regardless of political affiliation, have pleaded for this sort of reform.
AI CHATBOT PRIVACY: WHAT YOUR AI MAY KNOW ABOUT YOU
Mark Zuckerberg and the other innovators who built the social media revolution can rightly argue that they did not know, at least in the very beginning, what they were creating or the impact it would have. Nobody in a Harvard dorm room in 2004 had a body of evidence showing what Facebook would do to the sleep patterns, self-esteem, attention spans, friendships, politics, and mental health of the children coming of age behind them.
But the people who are now building and selling AI have no such excuse.
They have the entire social media experience sitting right in front of them. They know what happens when extraordinary technology meets the human desire for attention, companionship, affirmation, stimulation, and convenience. They apprehend the commercial incentive to keep the customer engaged for another minute, another hour, another year.
And AI presents a problem that social media never did.
MAJORITY OF YOUNG ADULTS SAY THEY ARE 'MORE CONCERNED THAN EXCITED' ABOUT AI: PEW RESEARCH
Facebook can influence your child. Instagram can make your child feel inadequate. TikTok can keep your child staring at a screen until two in the morning.
AI can become your child’s friend.
Or tutor, therapist, confidant, or adviser. Or romantic interest. Or some combination of all of these relationships.
It can know what your child is afraid of. Who your child has a crush on. What your child hates about their body. What is going wrong at home. What your child hasn’t told a parent, a teacher, a sibling, or a best friend. A generation of children could grow up pouring the most intimate details of their lives into machines owned and operated by corporations.
Think about that for a moment.
The danger isn’t just what AI tells our children. It is what our children tell AI.
So far, most AI executives, with a technological power that makes the social media platforms look like humble floppy disks from Radio Shack, have not shown much interest in doing the right thing.
But they should. They should take a look around at their communities, their own children, and consider the brave new world they are poised to create.
There are obvious places to begin. Parents should know when and how their children are using AI. Children should always know when they are dealing with a machine rather than a human being. There should be serious limits on the collection and retention of children’s most personal information. AI companionship for minors deserves scrutiny of an entirely different order than asking a chatbot who won the Battle of Gettysburg. There should be limits on how aggressively these products can be designed to keep a child engaged, and how they can be used, particularly late at night and without a parent’s knowledge.
The major AI companies could agree on a set of basic rules for children, an AI Geneva Convention of sorts. Not because these companies have suddenly become charities, and not because competition has ceased to matter. As Meta urged TikTok and YouTube, if everybody plays by the same rules, nobody is forced to choose between responsible behavior and market share.
The standard Silicon Valley response to regulation has been that anything that slows American technology hands the future to China. There is truth in that. The United States cannot unilaterally disarm in the AI race while Beijing charges forward.
But protecting American children is not unilateral disarmament.
CLICK HERE FOR MORE FOX NEWS OPINION
We can race China to build the smartest AI without racing China to see who can most effectively manipulate a 12-year-old. We can lead the world in artificial intelligence without surrendering an entire generation to it.
For AI products designed specifically to interact intimately and repeatedly with children, perhaps the companies first should have to demonstrate that reasonable protections exist. The consequences of being too careful are only that a child waits a little longer for some new feature. The consequences of being too cavalier are potentially disastrous.
There is a still deeper question that doesn’t fit neatly into legislation or parental controls.
What happens to childhood when a child never has to be bored?
Or lonely. Or confused. Or stumped. Or forced to figure something out without assistance. What happens when every question receives an immediate answer, every insecurity an immediate reassurance, every empty hour an available companion?
Those things sound like problems technology should solve.
But boredom, uncertainty, frustration, loneliness, and even failure have been part of growing up since there have been children. These human challenges help produce imagination, independence, resilience, judgment, and the ability to exist inside your own head.
The strangest danger of AI may turn out not to be that it does terrible things for our children.
It may be that it does wonderful things for them, all the time, so easily and so completely that we discover too late that AI has eliminated essential parts of how children learn to become adults.
There surely will be casualties of the AI revolution. But our children should not be among them.