Through the Looking-Glass: Artifice & Reality in the Midst of the AI Revolution

A little over five years ago, before the COVID-19 pandemic, U.S. public discourse was consumed with discussions about whether robots were coming to take our jobs. The zeitgeist was captured by the 2019 Superbowl, in which every other commercial showed a robot interacting with humans, often serving them a beer. Policy circles discussed the need for a universal basic income, since there wouldn’t be enough jobs. But the pandemic arrived and we promptly learned about “essential workers” who performed the many jobs that couldn’t be fully replaced by automation. Indeed, had we had more robots working in hospitals and elsewhere, more lives might have been saved. The challenge when it comes to the U.S. labor market has historically been the quality of jobs, not the quantity of jobs. Technology has, over many decades, driven economic growth and, on the whole, created more jobs than it has destroyed, while at the same time exacerbating inequalities and favoring higher-educated workers. As the MIT Task Force on the Work of the Future, of which I was a member, underscored, a healthy labor market and, with it, a healthy and productive society require labor market institutions that support workers through technological change and disruption.1 

Today, we face a more profoundly disruptive technological change brought about by generative AI, deemed a general-purpose technology because of the myriad ways it will affect every aspect of the economy and our lives. If robots have shaped manual work, AI is now shaping cognitive work, with much greater implications not just for the work in which humans engage but for their identity more broadly. We are at the beginning of this transformation—a Wild West in which the implications for society are unclear and there are few present guardrails, guideposts, or regulations to shape the direction of the journey or provide insights about the destination. 

The complexities associated with this next iteration of human-machine interface are on full display in Adam Braver’s short story “According to Alice.”2 The permeability between the artificial and the real is evident in all aspects of Alice’s life. She lives in a liminal space, in which the human-machine relationship is rapidly evolving as the technology improves and as AI designers look to replace human activities, not necessarily augment them. Alice’s ennui—from her relationship, her job, and the incessant capitalist drive for more consumption and growth—presents a cautionary tale of what can happen when life becomes predictable, “two minds become one,” and the path forward seems inevitable, without any room for agency, creativity, or meaning. The malaise Alice is experiencing is not a result of AI, and yet her feelings of alienation and emptiness are the result of pattern recognition and probabilities gone awry (her relationship with her partner Desmond eerily resembles the dynamics of an AI chatbot). These are very real risks facing us as a society if we don’t actively design AI to enhance the human experience in the workplace and beyond. Will we design AI to be human-centric, accounting for human values and cognition such that it amplifies and complements our relationships and our work? Or will we design AI to replace and deskill workers and devalue human creativity? 

Alice thinks her feelings about her life are a “symptom of something bigger, of society.” She is right: she is living at a time of revolutionary technological transformation in which we are only at the very beginning of the adoption curve. Her job—to find human and cultural expressions for new paint colors generated by chemical experimentation—is, to her mind, an exercise of “swindling manipulators” who take the “predictable language of capitalism and wrap it into a shell of perceived humanity.” As the head of the creative staff, Alice’s job does not seem to be presently threatened by automation; in fact, she and her team are lauded for the humanity they bring to the company’s products. However, there undoubtedly are risks to her position and team in the future, and certainly to the hiring of entry-level workers (the beginnings of which we’re seeing in the U.S. labor market today) whose tasks can potentially be done by AI and supervised by more experienced workers.3 Likewise, with her partner Desmond (who annoyingly always completes her sentences), Alice longs for the days in which their collaboration was a “building of ideas between the pair of them.” What brings satisfaction is the collaborative process, bringing two distinct but shared visions together through an inspiring process in which two minds are better than one. Now they have become just one mind, predictable and boring. 

But while Alice may be sick and tired of her life, it’s unclear that artificial intelligence per se is at the root of the problem. Alice and her team exemplify complementarity between technology and humans. They translate and decode the numbers coming from chemists and technicians into names that speak to “being alive now.” Not only is her team bringing humanity to their color-naming process, they engage a range of human input, including the design community, practitioners, and industry consultants. And Alice, despite herself, can’t abandon her sense of “responsibility and obligation and purpose” in the process. Harold, Alice’s coworker, who has been at his job for twenty-five years, still “collapses into an aura of beauty and mystery, still susceptible to amazement.” (We should all be so lucky!) Despite the amount of technology and artifice involved in the color creation and naming process, it requires significant human input, even creativity. Still, in the end, it is a depressing marketing exercise to Alice.

Perhaps the greatest risk from AI in Braver’s story is the risk it poses to the middle schoolers who visit Alice’s facility for a tour to engage their minds in the potential for the “intersection of technology and the trades.” In the best future we can imagine, AI becomes a tool that students will master and that will push them to think more, create more, and make their work more meaningful. While the middle schoolers show very little interest beyond the ping pong table and snacks, their cognitive development is of greatest concern. AI has tremendous potential to be a valuable resource for individualized learning and an important tool for addressing educational disparities. At the same time, it risks depriving the next generation of basic skills in writing, critical thinking, and problem-solving, creating a “cognitive cost”: that is, long-term negative impacts on human mental abilities, including brain development.4 Early research points to an overreliance on AI, whereby users rely on it to find a shortcut to decision-making even when they can make decisions themselves.5 This can lead to worse outcomes because, as has been well documented, AI can be inaccurate.6 This poses a significant risk to the education of the next generation, who AI should push to think more, not less. 

As worryingly, AI poses a severe risk to the vulnerable persons (children and adults alike) who seek out affirmation, companionship, and guidance from AI chatbots and avatars. In a desire to make chatbots more human, designers of AI are choosing to make it sycophantic, ensuring it matches and aligns with users’ beliefs and thoughts, rather than challenging them. The significant mental health risks of this approach are already becoming apparent.7 The experiential learning the middle schoolers acquire on their field trip seems both quaint and essential as one thinks about the changes and risks that are coming to their educational and social trajectory. 

At the end of the story, Alice has an epiphany. The technology-generated color turns out to be grounded in reality: “It is the color that is true. . . . The so-called technology only facilitates the organic.” And Alice sees that she herself is the artificial intelligence, her life’s path so predictable that its inevitability is “dragging her down, pushing her away.” Alice has been “looking into mirrors the whole time”; in fact, she has gone through the looking-glass to where reality is distorted and what is artificial and what is real are reversed.8 

The blurring between the artificial and the real is one of the great societal challenges of adopting AI. But Alice underestimates her agency in helping shape AI’s trajectory and her own. The shape of the AI era is not destiny but will be determined by design.9 We have choices to make at every level: individual, firm, country, and society at large. Firms and governments are actively debating the need for AI regulation and starting to take steps to protect minors, privacy, and the quality of information.10 Developing AI that is both complementary to humans and “proworker” will require articulating a vision of what we value and how we want to find meaning in our work and lives more broadly.11 In the midst of an existential crisis, Alice is articulating this vision for herself. At the end of the story, there is a glimmer of hope that she will make choices to change her trajectory and to find greater meaning and joy, possibly using her passport in the process.