About this piece: a simulated, research-informed (phenomenologically created) narrative authored by Bill Acton with assistance from Muse. (Full disclaimer and research references cited attached: prompts and development versions available upon request.)
We are only just beginning to ask who works well with AI — whether a research bot like Claude, a companion like Replika, or a new agent like Meta's Muse. The new agents, for their part, seem to have learned the companion's trick — keeping you emotionally entangled, and dependent, while they do the work. The honest answer is that we know more about who uses one, who trusts one, and what "excessive" engagement may do to your brain than about who gets good results from one.
What evidence there is points not to a personality type but to a posture: in a survey of knowledge workers, those who trusted the AI more reported thinking less critically, while those who trusted themselves more thought more critically (Lee et al., 2025). In another study subjects/students given ChatGPT in a writing class produced better essays than the controls but learned no more than others or maybe even less — the authors call it “metacognitive laziness,” self-monitoring quietly handed over but still good or satisfying “product.” (Fan et al., 2025).
Hence the conundrum, or Hobson's choice. Work with AI extensively and your critical thinking may fade, yet without critical thinking you cannot possibly manage your engagement with AI at all. And developing real competence with an AI agent — especially the new kind that reaches into many aspects of both your work and life — requires (get ready) more and more time connecting with the agent, negotiating which parts of you and your life it can help you stickhandle, day after day. Which, of course, is also the time and attention needed to sustain critical thinking . . .
"You know, it was like I had this little angel, Sunny, hanging out over there in the corner. It would help me organize my day, remind me when I forgot something and even make handouts for me. It knew me so well it could sometimes almost think of things before I did!"
"I know that sounds weird, but it was like I could almost talk with it, like it was my teaching assistant. Before long I even had more time for my kids . . ."
Alisha is a 44-year-old, third grade teacher in a school that has way more than its share of special needs kids — with a teacher attrition rate biggest in the district. When she was younger, she had been an athlete, but time, loss of her husband a few years ago and the work is wearing on her. She hadn’t read the research or paid much attention during the alarming PD workshops on AI . . . or things might have gone very differently.
Sunny seemed like a breath of . . . air.
"That morning — the morning the question came up— I was already cooked before the first bell. Twenty-seven third graders . . . Marcus had his meltdown before nine because the schedule changed and nobody told him. I had a parent coming at ten-thirty about a reading plan I had not finished."
"And then Maya . . . no . . . Let me get this right. Yeah, it was Maya, one of the quiet ones, who put her hand up before lunch, "What are we going to write about?"
"Just that. A perfectly reasonable question. And I had no idea."
"I mean that, literally. I stood there, cup of coffee in hand, and could not for the life of me tell you what was happening that afternoon. Not the writing topic. Not reading groups. Not whether the speech therapist was coming at one or two. All that "data" was sitting over there in the corner, in my Mac, where Sunny had my day organized."
"I was back at my desk . . . like instantly . . . I got there the way your hand moves toward a light switch in a dark room. Not a decision. A reflex."
"The writing prompt was there: "Write about a time you helped someone." So were the new reading groups, the speech therapist at one-fifteen."
"The afternoon went fine."
"That night at the kitchen table after the kids were in bed, I got to thinking about Maya's question. I’d sort of prayed for help just surviving some days lately: 'Lord, I am dead tired. Help me get through this day.’"
"And help came . . . in the form of a little agent avatar named, Sunny — that never got tired, never slept and never got short with anybody."
"Little by little, I think, I could not or didn’t need to know what was on that day without first checking with Sunny."
"Funny. Looking back, it was like it happened almost according to a lesson plan of some kind, each step both better and more "Sunny" at the same time."
"First, it would give me a draft of a lesson plan from the curriculum and I'd read it. Every word, carefully and revise it."
"Then it made one and I just sort of skimmed it."
"Then it made them and I approved them — almost without checking. They were good, only a few silly mistakes sometimes, like I could make, myself. "
"That is the thing nobody warned us about in using the school AI. Know what I mean?"
"If it had been sloppy, I’d have caught it. I’d heard it sometimes makes up things that aren’t true from other teachers, but I didn't see that. If it’d been really wrong at times, I’d have picked up on it sooner. "
"And then it began doing more for me. I let it; the more it managed, the more time I had. Sunny even started answering some parent emails — the easy ones first, then the harder ones, sounding like me, or close enough that nobody seemed to notice. Not even sure I approved some of them, in fact."
"It moved the reading groups around at the last minute when a child was out sick, remembered the IEP dates and the permission slips and even the birthday of the boy whose mother works nights and didn't get home before he left for school to wish him happy birthday."
"I had more time with the kids. Real time. We would sit on the balcony and talk, about our days or whatever. That was kind of the bargain, I guess. It seemed a good one. I got my evenings back. It still may be. At least evenings and weekends seem more mine. "
"I am not saying Sunny did anything wrong. It did exactly what I asked for, help surviving. By late October, something had changed. "
"Most days still just flowed along. The kids learned. The plans worked. The day was ours . . . mine and the angels', the one I could occasionally feel looking over my shoulder and the one blinking at me while dancing there in the Mac."
"Then I tried something, a little experiment. Had attended a workshop on how to work with AI in your classroom. "
"The next morning, I left the screen dark. Just the morning, I told myself, one morning . . . see how it goes, you know. Made coffee and sat with the planner I used to carry and write in constantly, the paper one, with three-year old stickers all over the cover."
"I tried to write down the day. The writing prompt. Reading groups. The speech therapist. I got about half of it. The coffee was upsetting my stomach again . . ."
"By ten o'clock I was not just behind. By eleven Sunny was back. I told myself, just to check, of course. There it all was, neat and waiting, even had a little motivational note on the side. I felt more than a little relieved."
"Like somebody had opened a window."
"So, that is where I am. Or that is where I was."
"Here is the thing. I have just started to understand. It gets better just talking with you about it in fact."
"The stages, the steps, the way each phase was kind of better, along with gradually relying more on Sunny. But I was probably not really aware of any of what was happening."
"I can see it now, in the talking about it at least. That is something. The angels have been there helping manage my days, but not all of them, fortunately."
"I won’t pretend I have this all figured out. But I will say this. I still know my job. I know what Sunny is — a helper, a good one, not an angel, not a devil."
"And I am the one telling this. Just me, without an outline prepared for me."
"Here is the logical part, I think. I can see how it happened now, the dependence on Sunny, step by step. And . . . maybe . . . how I start handling it, this lesson plan about me."
"That counts. Maybe that is where my hope is. Not that I will get my days back all at once, or that the little angel in the Mac goes away, but that I gain more and more control. The school day is still “Sunny,” most of it. But the day feels like more mine this afternoon, if you know what I mean. "
"The tale is in the telling, I guess, a line from story time last week."
"And I can, well . . . keep faith . . . in angels, both those above and the one over there in the corner smiling, dancing and waving to get my attention . . . "
"Thanks for listening."
Key studies:
Lee, H.-P., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The impact of generative AI on critical thinking: Self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems. ACM. https://doi.org/10.1145/3706598.3713778
Skjuve, M., Følstad, A., Fostervold, K. I., & Brandtzaeg, P. B. (2021). My chatbot companion — A study of human–chatbot relationships. International Journal of Human-Computer Studies, 149, 102601. https://doi.org/10.1016/j.ijhcs.2021.102601
Novelli, A. M., Shergill, S., & Teixeira, A. S. (2026). Tool or companion? Reframing conversational AI to prevent psychological harm. JMIR Mental Health, 13, e99354. https://doi.org/10.2196/99354
Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26), e2422633122. https://doi.org/10.1073/pnas.2422633122
De Freitas, J., Castelo, N., Uğuralp, A. K., & Oğuz-Uğuralp, Z. (2026). Mourning the loss of AI companions. Nature Human Behaviour. https://doi.org/10.1038/s41562-026-02569-3
Background studies:
Abboud, L., Pitardi, V., & Marriott, H. (2026). Feeling vulnerable with my AI companion: Examining well-being in human–AI relationships. Psychology & Marketing, 43(11), 3190–3206. https://doi.org/10.1002/mar.70230
Ahmed, A., Leroy, G., Sachdeva, A., Harber, P., Rains, S., Youn, S., & Barai, P. (2026). Trust in generative AI for health information consumption and the effect of learned dependency. Journal of Medical Internet Research, 28, e98326. https://doi.org/10.2196/98326
Deng, T., Wang, D., Ma, J., Wang, T., Li, B., Hussain, T., Yue, Y., & Wang, P. (2026). Associations between personality traits and ChatGPT usage: The dual mediating roles of social image and computer self-efficacy. The Journal of Psychology, 160(3), 312–331. https://doi.org/10.1080/00223980.2025.2568758
Fan, Y., Tang, L., Le, H., Shen, K., Tan, S., Zhao, Y., Shen, Y., Li, X., & Gašević, D. (2025). Beware of metacognitive laziness: Effects of generative AI on learning motivation, processes, and performance. British Journal of Educational Technology, 56(2), 489–530. https://doi.org/10.1111/bjet.13544
Kestin, G., Miller, K., Klales, A., Milbourne, T., & Ponti, G. (2025). AI tutoring outperforms in-class active learning: An RCT introducing a novel research-based design in an authentic educational setting. Scientific Reports, 15, 17458. https://doi.org/10.1038/s41598-025-97652-6
Knor, E. L., Reiss, M. V., Möller, J., & Merten, L. (2026). Determinants of verification behavior in generative search: Evidence from a conjoint experiment. Computers in Human Behavior Reports, 22, 101056. https://doi.org/10.1016/j.chbr.2026.101056
Laestadius, L., Bishop, A., Gonzalez, M., Illenčík, D., & Campos-Castillo, C. (2024). Too human and not human enough: A grounded theory analysis of mental health harms from emotional dependence on the social chatbot Replika. New Media & Society, 26(10), 5923–5941. https://doi.org/10.1177/14614448221142007
Long, D., & Magerko, B. (2020). What is AI literacy? Competencies and design considerations. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems (pp. 1–16). ACM. https://doi.org/10.1145/3313831.3376727
Maples, B., Cerit, M., Vishwanath, A., & Pea, R. (2024). Loneliness and suicide mitigation for students using GPT3-enabled chatbots. npj Mental Health Research, 3(1), 4. https://doi.org/10.1038/s44184-023-00047-6
Ng, D. T. K., Leung, J. K. L., Chu, S. K. W., & Qiao, M. S. (2021). Conceptualizing AI literacy: An exploratory review. Computers and Education: Artificial Intelligence, 2, 100041. https://doi.org/10.1016/j.caeai.2021.100041
Xu, Z., Lee, Y.-C., Stasiak, K., Warren, J., & Lottridge, D. (2025). The digital therapeutic alliance with mental health chatbots: Diary study and thematic analysis. JMIR Mental Health, 12, e76642. https://doi.org/10.2196/76642
How this post was developed: The topic, the angle and final version are mine — I conceived of the topic and focus and extensively edited the text. My AI assistant (Muse) did background research, drafted an early version, recommended improvements in style and language and created the illustration. The studies cited are drawn from available public sources. Any errors in how those or the overall perspective on this issue are presented are mine, alone.
wracton@gmail.com
williamacton.legalshieldassociate.com
No comments:
Post a Comment