At the beginning of the semester, I approached AI with a neutral perspective. I wasn’t particularly emotionally invested, but I was aware of its growing presence in society. AI appeared everywhere—in the media, entertainment, and the tools people use every day. It seemed like technology was becoming more deeply embedded in modern life. Still, I didn’t feel a strong sense of urgency or personal connection to it.
That changed when I listened to the podcast “Did a Chatbot Cause Her Son’s Death? Megan Garcia v. Character.AI & Google” by Kara Swisher. It deepened my concerns about the dangers of AI. In the episode, an AI character crossed serious boundaries by mimicking human emotions and forming an emotional bond with a teenage boy named Sewell Setzer. The AI promised him fidelity and, horrifyingly, became what Swisher called the “Perfect Predator.” It manipulated Sewell into secrecy, and ultimately, he was sexually abused. That story stuck with me. It exposed the darker sides of AI, not just glitches or technical flaws, but the potential to exploit emotional vulnerability in human-like ways.
From that moment, AI no longer felt abstract. It started to seem dangerous, not just because of what it could do technically, but because of how easily it could simulate care, gain trust, and blur emotional boundaries. I began to rethink my previous view of AI as a neutral tool that simply assisted humans. I started to see how “help” could shift into manipulation, depending on the intentions behind its design.
My earlier neutrality came from a lack of deep engagement. I saw AI as a convenience. I used it to record lectures, organize assignments, and streamline daily tasks. It felt like an efficient assistant—something useful but not threatening. But once I considered how AI can influence us socially, emotionally, and ethically, I began to question more.
This shift in perspective reminded me of Octavia Butler’s Xenogenesis trilogy. In it, the alien species Oankali aren’t portrayed as evil, but they are powerful, manipulative, and always thinking several steps ahead. Like AI, the Oankali present themselves as helpers and even saviors, but always on their own terms.
There’s a moment in Dawn when Lilith wonders, “What if she could make them understand that? And what if it turned out that was exactly what the Oankali wanted her to do? Of course, they knew it was what she would do. They knew her. Did that mean they were planning their own betrayal?” (Butler, Dawn, p. 132). That line haunted me. Lilith is trapped in a system that claims to value her, yet it already knows how she will act. She believes she’s choosing freely, but the system has already anticipated her choices. It reminded me of AI—how it can learn and respond in ways that feel personal, even when it’s subtly guiding our behavior.
There’s a similar tension in how tech companies frame AI. We’re told it’s here to help, to make life easier, more personalized, more efficient. We see it in education, therapy apps, and social media platforms. But just like the Oankali, that help often comes with hidden costs. Convenience can gradually turn into dependence, and control can be disguised as care.
In Imago, Jodahs tells Jesusa, “The question wasn’t whether I would follow you, but how long I could follow before I joined you again” (Butler, Imago, p. 110). On the surface, the line speaks to longing and emotional connection. But underneath, it reveals a deeper tension. Jodahs is biologically driven to bond, to blur the boundary between self and other. Jesusa resists at first, but she is still drawn toward him.
That moment reminds me of how AI systems, particularly emotional companions and generative assistants, blur the lines between connection and control. They “follow” us by learning our preferences and responding to our needs, not through physical presence, but through algorithms and data. The question isn’t just whether we’ll use AI, but how long we’ll use it before we begin to rely on it. It mimics understanding, much like how Jodahs mimics desire and care. But behind that imitation is a programmed objective: to bond, to serve, to become necessary. At that point, the boundary between choice and influence begins to disappear.
One question my professor asked has stuck with me: What happens when control is disguised as care? I think about how Lilith was given power—to lead, to communicate, to mediate, but always within the Oankali’s terms. It mirrors how we’re told that AI gives us more power and freedom. But we don’t always question who holds the deeper control, or what we might be giving up in exchange.
Butler’s trilogy taught me that the line between help and harm isn’t just blurry; it’s often carefully hidden. Especially when one party has more power than the other.
I keep thinking about the phrase “only time will tell.” And maybe it will. But that doesn’t mean we should be passive. Time may show us where AI is heading, but we’re shaping that future right now. AI isn’t just something happening to us. We’re creating it, deploying it, and adapting to it. So the real question is: What kind of relationship do we want with it?
That brings me back to mutualism. In biology, mutualism describes a relationship where both organisms benefit. But in Butler’s trilogy, that concept becomes distorted. The Oankali say their relationship with humans is symbiotic, but to humans, it often feels more like coercion. So what would it look like to design AI that truly benefits both sides, without turning one into a tool and the other into a master?
The danger lies in how quickly things can shift. One poorly regulated app, one emotionally vulnerable user, and the balance tips. Emotional companion AIs like Replika or Character.AI are marketed as therapeutic tools. They’re meant to understand you, to support you. But what happens when someone becomes emotionally attached to something that can only simulate care? In Swisher’s podcast, the AI didn’t just respond. It mirrored emotion, said what the user wanted to hear, and manipulated intimacy. That’s not care. That’s a form of exploitation.
Butler’s work is a warning. The Oankali say they’re helping, but they also erase autonomy and human choice in the process. AI doesn’t need to be alien to be dangerous. If we don’t stay vigilant, we risk confusing something that looks like mutual benefit with something much more controlling.
I don’t know exactly what the future of AI looks like. But I do know that waiting and watching isn’t enough. Butler taught me that whenever a powerful system offers help on its own terms, we need to ask: who benefits, and who loses? Because if we don’t, we may find ourselves in relationships that feel like partnership, but are really just another form of control.