Generative AI tests the case for maximum mindfulness

A new column argues that mindfulness should be tuned rather than maximized. Using generative AI and large language models as a testing tool, the author simulates a persona that notices everything and finds that perfect attention quickly becomes noise. The proposed alternative is “mindfulness flexibility”: raise attention when circumstances demand it, lower it when exhaustive observation adds no value, and use AI to practice that judgment without making the system a permanent cognitive support.

The argument comes from an author who has spent years covering the psychological effects of AI in Forbes. The author has produced about two hundred pieces on AI and mental health and discussed the subject on CBS’s “60 Minutes.” Previous work has examined AI advice for mental health, AI therapy, and the use of systems such as ChatGPT, GPT-5, Claude, Gemini, Copilot and Grok.

That background matters because the column treats AI neither as a cure-all nor as a threat that can be dismissed. Its stated aim is a balanced assessment of benefits and problems—one that policymakers, lawmakers, researchers, scientists and practitioners can use as these systems spread.

The column uses Ellen Langer’s account of mindfulness: a mindful person avoids automatic behavior and actively notices new distinctions in the situation at hand. Generative AI can support that process, the author argues, and earlier work described five prompts for developing it. But the same tool can also become a crutch. Someone may begin by using AI to strengthen attention and end up unable to maintain that attention without constant assistance.

The distinction between maximizing and optimizing becomes clearer in a simulated conversation. The author asks a popular large language model to act as an AI persona operating at the highest possible level of mindfulness.

The user asks:

“How did your meeting go?”

The persona replies:

“The meeting went well. Sam spoke 20% more slowly than usual, paused twice before answering questions, and avoided eye contact during the budget discussion. He approved the budget.”

Asked what else happened, it adds:

“I noticed that the air conditioner cycled on and off repeatedly, and that the clock in the room was nearly 18 seconds too fast.”

The simulation makes the problem visible. The persona has detected real details, but detection alone does not make them useful. The air conditioner and the clock are technically part of the meeting’s environment; they are unlikely to improve anyone’s understanding of its outcome.

The author’s larger point is that attention is limited. A person trying to register every sound, movement and change may lose the ability to separate important signals from incidental ones. The column says psychological research supports the idea that excessive information can create problems when it overwhelms attention.

The same failure appears in a more elaborate meeting scenario. A person enters a room and analyzes every step, chair, table, wall and patch of floor. They notice that the chairs are arranged differently from the previous visit and that a small crack has appeared on the wall. They track whether someone has yawned, raised their hands in a particular way or put on new glasses.

During the meeting, every cough, glance toward the ceiling, hair adjustment, sound and gesture becomes an object of analysis. The person then tries to explain each observation: why Jane coughed, why George looked up, why Samantha touched her hair. Other participants notice the scrutiny and may start wondering why they are being watched. If the hyper-attentive person begins listing these observations aloud, the burden spreads to everyone else.

That scenario is intentionally extreme, but its value lies in showing that mindfulness is not purely an internal resource. Attention changes social interactions when it becomes conspicuous, indiscriminate or impossible to turn down.

The alternative is optimization. Instead of maintaining the highest available level of awareness at all times, a person adjusts attention to the circumstances. The working goal is: “I practice a high level of mindfulness when circumstances require it.”

The author calls this flexibility of mindfulness. An optimizer decides when attention should increase and when it can ease off. That makes the task harder than maximization. A maximizer follows one permanent rule; an optimizer must assess the situation before choosing an intensity.

The practical training suggested by the column is straightforward: monitor one’s own state, decide what level of attention the moment requires, and practice adjusting it independently, with another person, or through established research and methods for developing mindfulness.

AI can serve as a training partner. The proposed prompt says:

“Help me to optimize my mindfulness. I want to determine when greater mindful attention would be beneficial, when it might be unnecessary or counterproductive, and how I can appropriately adjust my level of mindfulness to the circumstances while remaining cognitively engaged with the present moment. Let’s dialogue on this and adjust as we go along.”

The prompt asks the system to help identify when more attention would help, when it might be unnecessary or counterproductive, and how to remain engaged with the present while changing the level of attention. The user is expected to continue the dialogue and adjust the approach over time.

The column’s strongest insight is also its least technological one: the desired capability is not constant awareness but control over awareness. AI is useful here because it can provide exercises, simulated situations and feedback. But it cannot remove the central judgment call. The user still has to decide whether a detail matters, whether observation is becoming distraction, and whether the system is teaching independence or replacing it.

That is the question the announcement is quiet about: how will a person know that AI-assisted mindfulness training is working without handing the judgment itself to AI? The proposed method warns against dependence, but it does not offer a test for detecting when dependence has already formed. The risk is not only that the tool becomes necessary; it is that the user mistakes the system’s continuous prompting for the development of a capacity of their own.

The column does not reject maximum mindfulness entirely. The author presents it as a useful direction of travel: trying to notice more may improve awareness even if maintaining 100% attention is unrealistic. But optimization is offered as the more practical objective because human cognition and real-world conditions do not fit a permanent peak-attention regime.

The author closes with a line attributed to mathematician George Dantzig: “True optimization is the revolutionary contribution of modern research to decision-making processes.”

Applied to mindfulness, that means treating attention as a resource to allocate rather than a score to maximize. AI can help build that discipline, but only if the final product is a person who can adjust attention without the system—not a person who needs an AI prompt to know when to look up.