Be Ambivalent About AI
Why Mixed Feelings Beat No Feelings
It's fun when I'm exposed to a different way of thinking about something I've thought a lot about. That happened recently when listening to psychologist Naomi Rothman from Lehigh University discuss the power of ambivalence on the Hidden Brain podcast.
Many people think of ambivalence as being indecisive or noncommittal, but the word really means "having mixed feelings or contradictory ideas about something." It’s not fence-sitting; it’s hopping from one side of the fence to the other. I’m going to focus on emotional ambivalence, where opposing emotions are encapsulated by words like bittersweet, poignant, wistful, conflicted, nostalgic, and ironic.
It relates to perhaps the central cognitive question of our era. Can we override with intellect and practice the natural human inclination to minimize nuance and complexity and contrasting points of view? As I said in my previous article The State of AI Mindset, Two Years Later, “We are in a transition from a black-and-white world to a complex and nuanced one, and most adult brains aren’t handling that well.”
As Dr. Rothman pointed out from her research, ambivalence helps us make better decisions. The ability to take multiple perspectives helps decision making. That has been known for a long time. But most views of rationality try to eliminate emotion from the decision. What I interpreted from the episode is that appropriately used emotions can improve a decision, not hurt it.
Maybe I’ve been addressing emotional AI reactions the wrong way. While being careful not to dismiss emotions, which there are good reasons to have, I’ve been trying to supply the rational middle ground, asking educators to put aside their emotions so both pros and cons can be addressed. Maybe I should be trying to get people to feel the emotion of both anti-AI and pro-AI sides instead.
Schooling is Anti-Ambivalent
A lot of schooling reinforces binary thinking. The most painful for me to watch is in STEM, an inherently creative realm that we’ve reduced to formulas, repetition, and emotionless logic. The ethos of most education is that there is a right answer, and that knowing a bunch of the right answers makes one an expert. What I wish is education that’s filled with lots of discomfort, since life is nuanced, uncertain, and filled with those pesky, unpredictable Homo Sapiens.
It's not a place our brains like, so getting comfortable with the discomfort is key. Emotional states need practice too.
Rothman's research shows that our brains naturally seek to eliminate the discomfort of mixed feelings. Our educational system amplifies this tendency with lots of "do's" and "don'ts" and single, precise answers that only consider one aspect of a real-world problem. Teachers are positioned as bearers of knowledge and truth, shielding students from nuance and uncertainty.
Being able to stay in the zone of discomfort is a big part of being a great learner. As Rothman found in her studies, when people are primed to feel mixed emotions, they become "more open and receptive to peer advice" and "more accurate in their estimates" about complex situations. Yet educational systems often push us away from this productive discomfort.
There’s a lot of work trying to get students to put aside their humanity and approach problems as stoics. Centuries of this have led to the view that academia can avoid emotional bias, albeit through constant vigilance.
But this clearly doesn’t work. The amount of biased research and teaching is still high. The antidote may not be eliminating emotion further, which has done great damage to real people who occupy fields in STEM at a minimum. The cure might be getting people to feel multiple emotions about something.
Why Ambivalence Helps
Feeling mixed emotions actually makes us smarter decision-makers. When Rothman’s study participants experienced ambivalence, they processed information more broadly and inclusively. "If you have multiple emotions," she explains, "the error in the bias can be canceled out and therefore it can make us more accurate in our predictions and judgments about the world."
The catch is that ambivalence is uncomfortable, and we face social pressure against it. Rothman shows people tend to view those who express ambivalence as "less dominant, less confident, and less assertive." But in cooperative contexts, when people are working together to solve complex problems, expressing ambivalence suggested better outcomes. It seems the discomfort of mixed feelings can push us to think more deeply and consider multiple perspectives.
Creating Space for Mixed Feelings
This matters deeply for how we approach AI in education. AI cannot be thought of in black-and-white terms. It's not a monolith. Each product and use case have different considerations. Always treating AI as useless because it sometimes is, or as a forever panacea because it once was useful, is a more comfortable mental place to be. But it's not a more accurate one.
Researchers suggest developing what they call a "paradox mindset", viewing challenges with a "both/and approach instead of an either/or approach." As Rothman found in her studies, this works best in cooperative rather than competitive contexts, when people are "sitting actually on the same side of the table."
Emotional ambivalence—though uncomfortable—is a powerful tool for making better decisions in complex situations. Rather than trying to eliminate mixed feelings about AI in education, we should create spaces where we can productively explore them together.
Next time you feel torn about an AI implementation decision, resist the urge to rush to certainty. Instead, try sitting with the ambivalence and inviting others to explore the tensions with you. The discomfort might just lead to better choices.