AI can automate motivational interviewing, but transforming desires requires sustained engagement, not single conversations. We're now testing this with longer interactions.
Paper:
sciencedirect.com/science/arti... /6

In a preregistered trial with 237 participants, it successfully increased motivational readiness, but only for 24 hours. Then effects disappeared and never influenced actual prosocial behavior. /5

Conrado built Intrinsic AI, a GPT-4 chatbot that conducts motivational interviewing to help people internalize their goals. The idea: transform "I should" into "I want to." /4

Conventional wisdom says successful people have strong self-control. But recent evidence shows something else: People high in trait self-control use self-control less frequently, perhaps because they experience fewer temptations to begin with. /3

The self-help industry will generate $50 billion this year, yet most interventions fail within weeks. Why? We think it's because we've been focusing on the wrong mechanism. /2

The intervention is brief, scalable, and doesn't require overhauling existing programs. Worth reading if you work in mental health, education, or student services. Paper in Psychological Medicine. /5

www.cambridge.org/core/journal...

The results: false self-diagnosis was cut in half immediately and eliminated entirely at one-week follow-up. A practical way to balance awareness—informing people while minimizing unintended harms. /4

But Dasha developed a solution: brief "nocebo education" that teaches how negative expectations can make us misinterpret normal experiences as disorder symptoms. Just 10 minutes added to the workshop. /3

We tested this with ADHD awareness. 215 healthy young adults attended a workshop. Result: false self-diagnosis doubled afterward, despite no changes in actual symptoms. These beliefs persisted for at least a week. /2

So while we eagerly align with high-effort norms, we resist low-effort norms that threaten our self-image.

Huge congrats to Emily on this paper, now in press at JEP: General.

Paper: osf.io/9ygq4/ /end

This challenges how we think about social influence. We assume people conform to whatever norm they observe, high effort or low effort.

But effort is moralized. We use it to signal competence and moral character. /4

But when people learn that others are slacking off? They don't follow suit. They keep exerting effort.

Even when we made low effort more acceptable, people still resisted the low-effort norm. /3

Across 12 studies (N=2,084), we found a striking asymmetry: when people learn that others consistently choose harder tasks, they step up and work harder too. /2

It would be wrong to say I'm against the idea of cognitive fatigue having an entirely metabolic basis (it's a simpler explanation!), but there are empirical and theoretical hurdles these theories must address before they're (re)adopted as working hypotheses. We don't need a sequel to ego-depletion