We all know people who do not believe that artificial intelligence can ever be good for humanity, and many of them understandably wish we could abandon the technology altogether.

The concerns these critics cite include the environmental impact of data centers, questions about lack of financial reimbursement to creators for content used to train AI systems, and worries about what will happen as people increasingly turn to thinking machines to help them with cognitive tasks.

I respect people who wrestle with these issues, including those who desire to never have to engage with AI. However, as someone who works in the tech field, I am not focused on the question of whether people should use AI. I understand that that ship has already sailed and that we are long past the point of having the choice to opt out of AI technology. AI is already affecting all our lives, every day. The question I engage with therefore is not “Should I choose to use AI?” but rather “What do I need to know about how AI is showing up in all our lives, and how can I—and how can all of us—help to ensure that we are moving into the future with systems of AI that are serving us all well?”

The fact is, as a society, we’re at a tipping point: Will we give in to our fear of AI and cede control of it to the forces that are willing to engage with it? Or will we trade our fear for serious engagement and use our voices to ensure that we avoid the bad and capitalize on the good?

As an expert in the field of AI and robotics, I argue that the first option isn’t an option. Artificial intelligence is a reality, and if we are to thrive, not just survive, in an AI-infused world, we must take stock of how it is influencing our lives. The flaws inherent in AI should spur us all to speak out and to take action. Socrates once said that “the unexamined life is not worth living.” I propose a maxim that’s even more urgent: Unexamined AI is not only not worth having, it’s dangerous.

Ironically, the very qualities found in AI that could make an imaginary killer robot overlord possible—such as the ability to think intelligently, make decisions, and act without human assistance—are the same ones robots need to possess if they are to make our lives easier and maybe even better. That is to say, the same forces that could bring us HAL 9000, the Borg, and the Sentinels could also bring us R2-D2, Seven of Nine, and WALL-E.

The truth is that while the spread of AI presents unsettling risks, it offers us great promise too. The great technological shift we’re experiencing is not just taking old ways of being from us; it is also giving us all an opportunity to reexamine our lives and our values, and the values we share with our machines and algorithms. The result could be a richer, more rewarding future for humans and machines. But we’re going to have to work like hell, and fast, to get it right. Thankfully, we have what the movies might call a secret weapon: our humanness. That humanness is also, however, one of our greatest areas of vulnerability. Whichever path this quality leads us on, one thing is true. The future of AI comes down to one primary factor: people.

TRUST ME, I’M PRACTICALLY HUMAN

Before exploring the question of how our humanness will shape our future with AI, we must first understand how our instinct to trust influences that relationship. A unique aspect of my research, which differs from the work of other AI researchers, comes from studying what I call “overtrust”: that is, the human tendency to confidently trust technology without taking into account its limitations.

My curiosity about this phenomenon was sparked by my first research project as a PhD student at the University of Southern California, published in 1996, in which I proposed a robotic system to assist humans in the hospital environment by handling hazardous waste from infectious patients.¹ Although the theoretical underpinnings of this research prototype were crude, the project sparked my interest to explore the relationship between human workers and machines, and to examine what would persuade humans to trust machines as coworkers. Given the cultural forces we’ve already discussed, I expected to find people largely mistrustful of their robot/AI colleagues. I was surprised, then, when I discovered over the years that people can in fact be prone to trusting a robot, or an algorithm, too easily, a tendency that can lead to unintended consequences.

Now, some decades later, I am a trained robot whisperer who has mastered the language of robots as well as humans. In my area of expertise, understanding both is key. Why? Because while AI is in many ways unlike all other technologies that came before it, one of the most significant differences is that it’s driven by our relationships not with machines but with other people. Neuroscientific research shows that our brains have been hardwired to trust others. And those others don’t have to be human.

As it turns out, when we interact with AI, that contact triggers the same trusting feelings we experience in human-to-human relationships. For example, we’re okay with robots lying to us—as long as they are helpful.² We have no problem with divulging sensitive personal information to a robot—even when we know it makes mistakes.³ We are also averse to punishing robots—even if they’ve wronged us,⁴ indicating that we generally have a positive interpretation of their behavior.

And this is the point where my concern deepens. Under normal circumstances, identifying a problem leads to the opportunity to eliminate the complicating factor. But there is no way to eliminate the factor of our humanness, nor should we want to even if we could. Still, I worry about our ability to overcome this propensity to trust AI, given that it exploits the very element that makes us human.

And make no mistake: We are human and AI is not. Yet we often credit it with having a moral compass, just as humans do. But AI is neither inherently bad nor inherently good. It is only as ethical as it has been programmed to be. It is a reflection of us—that is, of those of us who program it and use it.

And that brings us back to the question of overtrust. Not only are we largely inclined to trust AI, but we also generally assume it has been programmed with our best interests in mind. AI developers, however, are not trained ethicists, and they don’t always think about the consequences of their creations. It’s up to us, then—everyday users of AI—to think more deeply about what we value as humans and what we want to accept as good (or good enough) choices in technology’s interactions with us, and to educate ourselves about the unintended consequences it has on the moral fabric of our society. With knowledge comes power, and understanding our triggers is one way to control AI instead of it being used to control us.

Another way to think about this is to say that AI doesn’t have an independent conscience; we are its conscience. Just because a company builds an AI solution doesn’t mean that we should use it in the first place or that we have to use it at all. As AI evolves and becomes more intelligent, it’s important for us to question its power and its use. Companies pushing AI solutions that we then blindly embrace can do us more harm than good. This dynamic is likely to worsen growing inequities in society and reinforce or introduce biases into people’s mindsets and in our communities. This is a deeper problem, a human issue.

ADVICE FROM AN AI OPTIMIST (AND REALIST)

Given all this doom-and-gloom talk, you may be surprised to learn that despite my expertise in the biases and dangers of AI, I do not view AI through the lens of a skeptic. In fact, I’ve lived with a utopian view of AI for a long time. I love working with this technology and have been doing so my entire adult life—and even before that. I programmed my first computer in the third grade and my first robot in the ninth. At the age of eighteen, I began working at NASA’s Jet Propulsion Laboratory and have been hooked on robots ever since.

My optimism has not led me to the margins of my field. Quite the contrary: My AI and robotics research has been supported over the past thirty years by NASA, Intel, the Grammy Foundation, Google, Microsoft, Amazon, and many other institutions. I regularly consult for and have sat on the advisory boards of leading AI and government organizations, such as the Association for the Advancement of Artificial Intelligence and National Artificial Intelligence Advisory Committee tasked to advise the US president. Over the years, I’ve discussed robotics over dinner with Jeff Bezos. I’ve engaged in discourse about AI, bias, and women in technology over breakfast with Melinda Gates, the cofounder of the Bill & Melinda Gates Foundation. I’ve conversed with Satya Nadella, CEO of Microsoft, about the role technology must play in making the world more accessible to everyone, including those with disabilities. In 2004, Time magazine featured me as an AI visionary and called me “The Bionic Engineer Driving School on Mars.” Business Insider has labeled me as one of the twenty-three most powerful women engineers in the world, and i magazine gave me a thumbs-up with a Top 50 US Women in Technology recognition. In between, I was even tagged by the Georgia Institute of Technology as the most senior Black woman roboticist alive, although I still feel pretty young for such a label.⁵

I tell you all this not to impress but to give you a sense of who I am and why you can trust me as your guide through the complexities of AI and its impact on our lives. Thirty years as a working roboticist has given me a lot of insights about this field, particularly because of my area of research and expertise, which is designing AI-driven robots that interact with humans. Some of these are science-driven robots (for exploring Mars and Earth), some are robots designed to give us improved health (such as those used by people with disabilities), and others are robots used in education (to improve our ability to learn).

One necessity in working in this area of expertise is that I have had to learn to understand human behavior in order to figure out what makes us humans tick—not because I actually care about humans (although I do) but because I need to help my robots learn how to personalize their interactions.

Within my work involving interactions between AI-driven robots and humans, I have still one more area of specialty: autonomy—that is, the use of methods in artificial intelligence to make my robots smarter. AI provides the intelligence to my machines, while autonomy enables those machines to learn and act independently using that intelligence. I realized early on in my career that attempts to create intelligent systems without giving them the ability to learn from their human counterparts were a dead end. I needed the robots to not just learn how to perform tasks; I needed them to do their work safely, in collaboration with people. My motivation was simple: I’d grown up grounded in a set of principles about right and wrong, and intentionally making robots that could hurt people was wrong. From that point on, I focused on laying the groundwork for—and then eventually making— robots that could learn to make human-like decisions.

Of course, when I started this work in the 1990s, the human–robot relationship was still a oneway street. Back then, robots were tools, not partners. But the robot world has since evolved. Today, AI learns by modeling how humans perform tasks and then uses this learned skill to perfect AI-driven robots’ abilities. No longer do humans have to actively try to teach robots to act like humans; the technology now lets us humans sit back and do all the things that humans do while AI does the work of figuring us out.

All of that sounds pretty good, doesn’t it? Unfortunately, over the past several years I’ve observed a subtle shift in the technology, one that’s driven by the idea that anything goes and by a growing human apathy and overreliance on technology. It is this shift that makes me believe that the scales of AI are shifting toward the dystopic. Make no mistake: I still fundamentally believe in the good of AI. But I’ve also come to a new understanding: The AI utopia I’ve envisioned for so many years is no longer possible without you.

BRIDGING THE GAP FOR EVERYONE

We as a society must ask numerous questions as we confront today’s problems and the realities surrounding AI. But underneath them all, the primary question underlying my current way of thinking is this: How do we ensure that AI makes the world better for all of us, and not just a select few? Because if we don’t turn the tide, that privilege of the select is where we’re headed. Trust me. Nobody knows better than a roboticist how very real that risk—far more urgent and pressing than any of Hollywood’s fantasies—truly is.

For the most part, this possible outcome isn’t the intended goal for AI developers. Some of the companies that build the AI we use on an everyday basis aren’t always explicitly aware of the harm they can create. Others don’t set out to cause harm but simply turn a blind eye to the problems.

And then there’s the other side of the AI coin: those who use it. Many leaders at companies, both large and small, feel a sense of pressure to adopt AI and do so without thinking through the why or how of that decision. Meanwhile, individual humans keep demanding more efficiency and more personalized services from the marketplace—and we get them, increasingly through the implementation of AI, though we frequently don’t even realize that’s what we’re using. As consumers, we eagerly give away our data in exchange for as little as a 5 percent off coupon, a new emoji, or a ten-minute reduction in commute time. We click “like” without thinking of what’s happening with our data—who’s profiting and how.

Companies that want to make money off of us take advantage of this, sweeping up our data and building models of our behavior, and spit new features at us. But these features aren’t simply intended to help us achieve faster commutes or hack our lives better; they’re intended to give companies more power over our lives and make it easier for them to snag and hold our attention. At the end of the day, their application of AI is about getting us to spend more money. As a result, they end up controlling not only our wallets, but our interactions with each other as well.

And this brings me back to my rising concern. Unfortunately, my observations have led me to fear that, despite not knowing how our data is being used, most of us will not use our power to say NO. Our privacy and data are becoming commoditized, while at the same time, the ways we perceive our world—the ways we humans understand one another, relate to one another, and work to make societies more or less functional—are being increasingly determined by the worldviews of the people who create these AI systems.

And here’s something else: Beyond fears of a robot apocalypse or job losses, there’s another reason to fear AI. It is increasingly in charge of running the world we live in, and it is as biased as humans can be. Remember, it learned from us. Bias doesn’t just influence what we do; it also affects what we see. And this too can prevent us from solving the problem at hand. Even those of us who consciously value diversity are often more comfortable with people who are like us. As a result, when we then see AI biased against others, we don’t always see it as an “us” problem, and so we don’t act to rectify the problem. I’m concerned that as we further categorize the world into “us” versus “them,” we will all fall into a state of apathy and not care until AI disadvantages all of us. By then it will be too late.

Another reason to feel concern about AI is that it is subject to the rules of “garbage in, garbage out.” This can easily turn flawed data into decision-making mistakes and false truths. These algorithms are so complex that it’s nearly impossible for anyone who isn’t an expert to check all their results. As they are used more, there’s a growing chance that these errors might cancel out any benefits they offer.

As you can see, the possibility of robot overlords isn’t our most immediate problem, though the problems we do face today threaten to oppress us in various ways. And we will need to make some real changes if we—as a society and as individuals—are going to get our power back.

THE QUESTIONS WE SHOULD BE ASKING

People continue to ask the wrong questions about AI. Consider common questions in the debate over the risks and rewards of self-driving cars. Do self-driving cars save lives? Yes. Will self-driving cars take away jobs? Also yes. But those questions assume that we humans are pawns in an AI game of chess and have no choice but to accept this new technology as it is. A different question we really should be asking about self-driving cars is: With whose values are these cars being programmed? Because they are being programmed with someone’s values. What’s often unclear is whether those values are yours.

Whenever you use a new AI-driven product, it’s safe for you to assume that you’re giving something up for that privilege. Ask yourself if what you’re getting is worth whatever you may be giving up. For instance, using facial recognition technology, China used AI to monitor its citizens extensively during the coronavirus crisis. Whether they were using the Close Contact Detector or Alipay Health Code app, citizens who adopted these tools would do well to ask themselves: Was the trade-off in civil liberties worth the purported benefit in public health?

The question of how AI is being used to control humans is a moral issue. Many people are certainly starting to see it as such. Before his death in 2025, Pope Francis warned against the dangers of AI and urged the big tech companies to ensure AI does not lead to a world in which those who can afford the best AI triumph over the weaker.⁶ Within the tech companies, some employees are becoming activists, trying to change the culture from within. Some have tried to force their management to become more transparent about how AI is used in their products or to limit whom it’s sold to. Others have signed petitions or written letters calling for a ban on their companies supplying AI for military use.

Even experts in the sector have become alarmed. In 2023, Geoff Hinton, considered one of the godfathers of AI, began a crusade to warn the public about this technology, most recently stating, “I often say [there’s a] 10% to 20% chance [for AI] to wipe us out. . . . And the hope is that if enough smart people do enough research with enough resources, we’ll figure out a way to build them so they’ll never want to harm us.”⁷ The most important part of that quote to me is the hope found in the “smart people”—and smart people include more than just the builders. Smart people include all of us.

While Hinton’s warning sounds alarmingly Hollywood-esque in tone, I take it as encouragement for all of us to be vigilant and stay engaged. My hope is not just that the days of robot overlords are in the distant future but also that they’re entirely avoidable.

AI’S BIG RED BUTTON

Although my work requires that I study human behaviors, which can be problematic at times, I am and continue to be a realistic optimist. I’m the type of person who will lend money to a relative who asks for it and say “pay it back when you can” while not really expecting to see it again.

When I programmed that first robot decades ago, I had no idea that AI would have such a vast impact on human lives. I also could not have predicted that my lived experience as an African American woman in technology—which I would have to navigate as an anomaly in that field—would ultimately give me the power to recognize, diagnose, and understand AI’s most dangerous flaws. Whether those flaws are around the categories of facial recognition technology, AI assistants, or weaponized AI like cyberhacking and consumer manipulation, the fact that I see the world differently allows me to suggest solutions that a diversity of people on earth can exploit. And I do, in fact, have great hope that we—that is, all of us who are a part of that diversity—can become AI activists.

In the field of robotics, engineers and roboticists talk about the “Big Red Button,” which is symbolic language for a robot kill switch: a safety mechanism designed to immediately shut off a robot in emergency situations or when the robot behaves unpredictably. It enables humans to maintain control over the robot and prevent it from harming people and property. In fact, when industrial robots are installed for use on a factory floor, the engineers and techs responsible usually install an actual, large emergency stop button as well. And in many science fiction movies, there’s often a scene near the very end of the film in which the heroine or hero just has to reach-reach-reach for some kind of stop button of their own.

In the real world, designing a big red button for physical robots is easy, and using it is simple. Pressing it just means you stop power to the motors, the computer, and the sensors. Virtual AI, on the other hand, has no big red button designed into its functionality. This seems, to put it mildly, unwise. The problem strikes me as rather like designing a self-driving car and then remembering afterward that it probably should have some sort of analog brakes that can be applied to stop the car if the car is not responding to prevent an impending danger or crash.

I’ve got some ideas on how we can all use AI better, and how to build a better future with it. And I do mean all of us. I firmly believe that it’s up to each and every person—not just researchers, developers, and programmers—to ensure the future of AI ends up being a good one for humanity.

The future of AI is now; it is flooding our lives at an exponential rate. But it’s not a natural disaster; it’s our creation. It’s shaped by us and learns from our example. You and I are ultimately responsible for the impact that it will have on our societies. AI is us and AI is our future. It’s in our best interest to be armed and prepared to deal with it.

Ironically enough, what we need to arm ourselves with is not weapons, at least not physical weapons of mass destruction. Quite the opposite: It’s our humanity. We can fortify our humanity and our collective will so that they will protect us against an AI-infused world. And I mean all of us, including those who are different from us.

Humanity is the one thing that AI doesn’t have and that it can’t learn from us. I optimistically believe that if we open our minds and embrace the differences that make us human, we have a real chance of preserving our humanity in the age of AI. And living with that outcome is a whole lot better than any doom-and-gloom Hollywood ending. 

Adapted from Rebooting the Machines, published by Broadleaf Books. Copyright © 2026 by Ayanna Howard. All rights reserved.

About the Author

Dr. Ayanna Howard is an accomplished entrepreneur, educator, and international expert in robotics and AI, and the president of Spelman College. In her former leadership roles, she has served as dean of The Ohio State University College of Engineering, chair of the School of Interactive Computing at the Georgia Institute of Technology, and a deputy manager at NASA’s Jet Propulsion Laboratory. She has been recognized on top industry lists including Forbes’ “America’s Top 50 Women in Tech,” Business Insider’s “23 of the Most Powerful Women Engineers,” and Women Inc.’s “Most Influential Women Corporate Board Directors.” She has also been featured in the Amazon Prime series Luminaries and in the IMAX movie Superhuman Body.

All author proceeds from the sale of Rebooting Machines go to Spelman College scholarships.

Endnotes

1. Ayanna Howard and George A. Bekey, “Prototype System for Automated Sorting and Removal of Bags of Hazardous Waste,” Proceedings of the SPIE 2904, Intelligent Robots and Computer Vision XV: Algorithms, Techniques, Active Vision and Materials Handling, pp. 271–277, Boston, MA, November 1996.
2. Kantwon Rogers and Ayanna Howard, “When a Robot Tells You That It Can Lie,” IEEE International Conference on Advanced Robotics and Its Social Impacts (ARSO), pp. 1–7, Long Beach, CA, 2022.
3. Jin Xu and Ayanna Howard, “Pilot Study for Examining Human–Robot Trust in Healthcare Interventions Involving Sensitive Personal Information,” Rehabilitation Engineering and Technology Society of North America (RESNA) Annual Conference, New Orleans, LA, June 2017.
4. Kantwon Rogers and Ayanna Howard, “Looks Can Permit Deceiving: How Reward or Punishment Decisions Are Influenced by Robot Embodiment,” ACM/IEEE International Conference on Human–Robot Interaction, pp. 57–61, New York, NY, 2021.
5. Jason Maderer, “Women of Robotics,” April 7, 2017.
6. Andrea Carlo, “Religion Against the Machine: Pope Francis Takes on AI,” Euronews, April 20, 2023. 
7. Ashton Jackson and Tom Huddleston Jr., “There’s a ‘10% to 20% Chance’ That AI Will Displace Humans Completely, Says ‘Godfather’ of the
Technology,” CNBC, June 2025.