Quiet Erasure 3: Pas de deux

Artificial intelligence is becoming extraordinarily adept at recognizing human feelings and responding in ways calibrated to our needs and vulnerabilities. Pas de deux imagines a quieter erasure: not a machine replacing love outright, but the skills and habits it rewards—anticipation, clarification, emotional regulation, a ready response—following us home. What happens when two people who care deeply for one another, both professionally and personally very much immersed in AI use and influence, begin to bring less of

Clker.com

rough, spontaneous, unfinished truth of themselves into the room?

This fictional embodiment of the implications of current research was generated initially by Perplexity.ai, responding to the prompt in the Appendix.

 
Quiet Erasure 3: Pas de deux

“Because there is a deep yearning in human beings to be understood.”
—Yuval Noah Harari

“You asked it first?”
“I asked it because I was still at work. You were home.”
“I know. And you asked it about us?”
“I didn’t think of it that way.”
“I don’t think you did. Forget it.”
He nodded, though he stayed at the table.
She stood at the counter with her bag still on one shoulder. Her laptop was open beside the mail.
“Rough day. Papers. Parent meetings. Then department chair walks in, wants the curriculum notes “made clearer, more concise,” and an executive summary for him to read at the board meeting tomorrow. I spent hours on it with Claude trying to clean it up. It’s political. You know how it is. 
“And instead of having me look it over later, you just did it yourself.”
“I was trying not to bring work home.”
“But you sort of brought Claude home with you.”
She dropped her bag on the table.
“What do you mean by that?”
“At my office, I do it too. I have meetings all day where every hesitation costs something. I ask it for the risks, the objections, the version that will land. I walk into rooms already prepared for people to disagree with me.”
“That makes sense.”
“It does there.”
“But not here.”
“No.” He glanced at the laptop. “Me, too. You know, I know sometimes I even prepare the same way in my head, on the way home, before I talk with you . . .  instead of just listening first.”
Her face softened, then tightened again.
“You prepare for me?”
“Not like that. Not on purpose.”
“But you do.”
He nodded.
She sat down at the table. “Maybe I expect that.”
“Do what?”
“Maybe I just sidestep a problem so you won’t have to deal with me.”
“I don’t want to manage you.”
“I know you don’t.” Rubbing her thumb against the rim of her mug, she responded. “But you come home tired. I see it before you say anything. And I think: don’t give him one more thing. Don’t be confused. Don’t be impossible.”
“You are never impossible.”
“I hope not. I just… I want to be at least easy to love when you have already had a hard day.”
He sat down across from her.
“You do not have to do that.”
“I know you mean that.”
“I do.”
 “Some days I don’t come to you with things I can’t make sense of . . . “
He did not answer right away.
She looked up. “Say something.”
“I’m trying not to make it sound too neat.”
“Good.”
“It’s not easy . . .”
“I know.”
He paused briefly. “When did you start doing that?”
“I don’t know.”
“Try.”
She shook her head. “I really don’t know.”
“Okay.”
“You always say that when you are trying to be patient.”
“I am trying to be patient.”
“With me?”
“With myself, too.”
She almost smiled . . .
He said, “Maybe it happened slowly.”
“What did?”
“This.” He touched the edge of the laptop. “Us rehearsing what we’ll say before we’re  together.”
She stared at the screen, speaking slower . . . “At school, I’ll have students who cannot begin, keep asking for the right first sentence, right outline, the right way to say what they mean.”
“And you tell them?”
“I tell them they have to go for it, risk a bad first sentence.” She swallowed. “I tell them we can work with anything honest.”
“What?”
“Nothing.”
“Daniel.”
“I haven’t given you much room for that bad first sentence.”
She didn’t react.
“At work, that’s what I do. When someone brings a problem, you don’t just sit on it. You come up with a next step, make the numbers work, something they can maybe use, sooner than later.”
“And with me?”
“With you, to be honest, I often hear your first few words and then, it’s like I’m already working on what I’m going to say back before you finish.”
She looked down at their hands, all four in soft fists . . . 
“I know,” she said.
“I’m so sorry.”
“I know you are.”
“No. No. I mean I’m sorry that you’ve had to do that with me.”
Her eyes filled.
“That is exactly what it feels like.”
He reached across the table. 
She put her hand in his. 
For a while, neither of them spoke.
Then she said, “When I got home, I got this email from him. I was so pissed. Really angry. He took credit for my work, what I had spent months building.  I felt ashamed for being that angry. Maybe I was being vain, territorial, childish. I don’t know.”
“You weren’t.”
“See? There it is!”
“You are right. I did it again.”
“I know you want to defend me.”
“I do, I did.”
“And I needed you to know more before we try to figure out how I should respond.”
He nodded. 
“The son of a bitch took credit for it.”
“And you were steamed.”
“Yeah, absolutely.”
“And then been ashamed of it?”
“Yes. A little, probably, at first.”
He held her hand. “That sounds devastating.”
Her mouth trembled.
“Yes,” she said. “It was.”
He waited.
“Then . . . I gave the email to Claude, along with my immediate reaction and asked it whether I was . . .  overreacting and how to respond.”
“What did it say?”
“It said my feelings were valid. No surprise there! It listed several possible explanations for his actions, and helped draft a response that was calm and professional.”
“Did you send it?”
“No.”
“Why not?”
“I don’t know.”
He waited. 
 “I think I wanted you to see it first.”
He looked at her.
 “I revised it. I had reasons ready. I. had a plan ready to show you.” 
He looked at the laptop again. “I do that too, Anna.”
“With me?”
“Yes. After you said last week that I was never really off work, I asked it how people like me bring work home without ruining relationships.”
“You did?”
“I did.”
“What did it say?”
“A great deal, as you can imagine.”
“Of course.”
“You name it . . . habits. Schedules. Scripts for checking in. It even suggested I ask you how you prefer to receive support, if you can believe it.”
“That sounds helpful!”
He looked at her blankly. ““I came to dinner with questions ready.”
“I remember.”
“You asked me something, and I knew what I was supposed to say before you were done.”
“That’s what I mean.”
“Grok was at the table with us.”
She looked toward the window.
“No,” she said. “You weren’t.”
He flinched.
“I’m not saying it reflects on how you feel about me,” she said quickly. “I never doubt that.”
“I love you.”
“I love you, too.”
“No, I mean it.” His voice broke a little. “I need you to know that I am in. I am here, right here because I want you, our life. I want the life we have not even made yet.”
She pressed his hand against her cheek.
“I know,” she whispered.
“And I don’t want us so managed that nothing real gets through.”
She closed her eyes.
“I don’t want that either.”
“I don’t want you to become a problem I solve.”
“And I don’t want you to become someone I only bring things I’ve figured out.”
 “We are good at good ideas.”
“We are.”
“At work.”
“At work.”
“Maybe we need to stop bringing Claude and Grok to dinner.”
 “Not always. I don’t want to pretend it has no place. It helps me teach. It helps me when I have too much to do. Sometimes it helps me find a way in.”
“Me too.”
“But maybe not what we call on first.”
“Not first.”
“If I am hurt, I come hurt.”
“If I am confused, I come confused.”
“Before you ask it to tell you what I mean.”
“Before I ask it to tell me what you mean.”
“That one matters.”
“It does.”
“And if we need help?”
“We ask together.”
“Together?”
“But not before we’ve tried to understand each other.”

They looked at each other.

“Do you think we can do that?”
“I think we’ll forget sometimes.”
“I hope so.”
“I’ll still reach for the easy answer.”

“And I’ll still try to make myself easier.”
“We’ll be tired.”
“Yes.”
“But maybe we can notice sooner.”
“Notice what?”
“When it all gets too smooth.”

She smiled then, almost despite herself.
“I don’t want smooth.”
“Neither do I.”
“Stop me sometimes.”
“Not to fix you.”
“No. Just ask me if that’s really what I mean.”

He nodded. “And tell me when I start talking like I’ve already written the minutes.”

She laughed softly.
“I can do that.”
“Promise?”
“I can try.”
“That isn’t a promise.”
“It’s the only one I can make and mean.”

He smiled.

“Daniel?”
“Yes?”
“I was humiliated today.”
“I know.”

She shook her head.
 “No. You don’t know yet.”

He looked down. 
“Then tell me.”

He did not reach for an answer. 
He did not ask her to explain it before she could.

He waited.

And she began.

The initial prompt:

Let's take Harari's perspective on AI as point of departure for a third Quiet erasure phenomenological creative piece, integrating in the most current research studies that explore the impact of AI engagement on both intimate, one on one relationships and the fundamental importance of dialogic learning. For characters, use two professionals, a man and woman who are romantically very much connected but not yet married. They both use AI extensively at work and in their private lives and they are beginning to recognize what it is doing to the depth and intimacy of their relationship. They are struggling to figure out a way out but it is seemingly at same time becoming more difficult over time. This should be almost entirely dialogue, little commentary. The conversation concludes with them at least more dialogically connected, their only real hope grounded in their love for eachother. Try that. Keep the conversation about the same length as the previous two Quiet Erasure pieces.

wracton@gmail.com

www.williamacton.legalshieldassociate.com (to subscribe to Legalshield or IDshield)

The quiet erasure, continued: A mother's reckoning

A phenomenological sketch, companion piece to "The Quiet Erasure"

The “symptoms” dramatized here are not simply invented for literary effect.  In some sense this is an extension of the depiction in The Quiet Erasure, from personal/cognitive out to the interpersonal— the new, brief strange hesitation before responding to a child, the reflexive appeal to a chatbot during a disagreement with a spouse, a friend's fleeting reaction of “betrayal” when iPhone is consulted instead of her, the kids learning to outsource advice with their device, on their own conflicts. 

It is grounded in a growing body of empirical research on AI-mediated communication, "cognitive surrender," and erosion of trust when third-party (AI) is introduced into dialogue. (See the attached reference list.) AI arbitration in disputes can generate feelings of betrayal and diminished trust between partners; heavy AI reliance correlates with reduced tolerance for open-ended, unresolved conversation. Repeated outsourcing to AI may short-circuit the fundamental processes — friction, disagreement, shared reasoning — through which relationships and informed judgment evolve. 

Clker.com

This fictional embodiment of the implications of current research was generated initially by Perplexity.ai, responding to the prompt in the Appendix. 

***

She notices it first in the odd, unsettling pause before she responds to her daughter.

Emma, twelve, is talking about a friend who said something cruel at lunch, and there is a beat — just a beat — where instead of immediately reaching inward for the intuitive wisdom passed down to her through her genes, mothers, mentors and experience that she used to trust  her hand moves toward her phone. Not to look anything up, exactly. Just to . . . ask. See what it says. She catches herself sometimes but now not often enough . . . 

There used to be a simple “feel” to parenting three children spanning a decade of development — the five-year-old's tantrums, the ten-year-old's homework meltdowns, the fifteen-year-old's silences — the aches, anxieties and fears that were, she now understands, just the “stuff” of becoming and being a mother. She could sit, almost happily with not-knowing. She would fumble lovingly toward an answer with her husband late at night, the two of them turning a problem over and over like a stone, disagreeing, circling back, arriving somewhere. That friction was not a failure of parenting. It was parenting.

Now the chatbot often nails a solution before the disagreement can even properly begin. Her husband raises a concern about, say their son's screen time; she has already asked Grok before he finishes his thought--and reads the confident, bulleted response aloud . . . and watches something flicker across his face that she recognizes, accompanied by a dull twitch in her stomach, something like betrayal — the sense that she trusted a machine over the twelve years of shared understanding and judgment sitting across the table from her. He doesn't argue, he just goes quiet, or he says "sure, ask Grok" in a tone that isn't really agreement. She knows and she has stopped being able to stop it from happening

It has crept into her friendships too, the ones she used to think were immune, sacred even — the two women who have been for fifteen years, her closest confidantes. A coffee that once meandered for two unhurried hours through random, half-formed worries now clips along at a strange new tempo: someone raises a dilemma about an aging parent or a stalled marriage, and within a minute a phone is out,  an answer read aloud in the same bright, decisive register the machine always uses. The three of them used to be really good at not knowing together — that was, she realizes now, most of what the friendship was for. Lately she catches her own thumb moving toward the search bar mid-sentence, cutting off a friend who was still finding her way toward what she actually meant to say. She sees the same flicker she saw in her husband's face, quieter, more practiced, almost in hiding. Women who have known each other that long know when they are being uncoupled . . . 

The children get it differently. The five-year-old just wants her eyes; increasingly her eyes are elsewhere, mid-scroll, mid-query, present in body only. The ten-year-old has started saying "never mind" when she starts to answer a question with "let's ask—" before she even finishes what she'll ask. The fifteen-year-old has begun mirroring her: bringing conflicts not to her, but with an overheard comments on the phone to her friend like, ”Wait, wait, wait . .   my bot will know . . .” She senses the unspoken horror of watching your own worst habit reproduce itself in your baby . . . that she taught this. Not through instruction, but through simple, repeated demonstration: an open question is an intolerable thing to leave sitting in a room, on the table.

Even the dog knows. The old beagle, who has spent nine years reading the emotional weather of this house far better than any of them, has taken to sitting at her feet — phone lit, family voices trailing off unanswered in the next room — looking up at her with blank expression, ears low, eyes fixed on her face, asking a question no chatbot could ever answer. She wonders in passing what it says about her that a beagle's face is what finally makes her put down the damn phone.

She knows. It feels like a limb gone slightly numb — present, but slow to move. A strange, cold clarity is emerging of watching herself outsource not her competence as a mom but her most prized possession — her relationships themselves, one deferred, often trivial answer or judgment at a time. Are the people she loves most now quietly resigning to the fact that she is less and less fully in the house with them at all?  And what to do about it is a question that she and Grok and the beagle cannot answer.

***

References and sources consulted by Perplexity, in addition to those reference list already published with the first Quiet Erasure piece.

Gerlich, M. (2025, January). Increased AI use linked to eroding critical thinking skills. Phys.org. https://phys.org/news/2025-01-ai-linked-eroding-critical-skills.html

AI tools in society: Impacts on cognitive offloading and the future of critical thinking. (2025). AIGL Blog. https://www.aigl.blog/ai-tools-in-society-impacts-on-cognitive-offloading-and-the-future-of-critical-thinking/

Cognitive offloading, critical thinking and attitudes towards artificial intelligence. (2025). PubMed. https://pubmed.ncbi.nlm.nih.gov/42377671/

Talking and thinking with AI: How AI chatbots restructure epistemic identity and virtue. (2025). Poligrafi. https://ojs.zrs-kp.si/index.php/poligrafi/article/view/510

Talking and thinking with AI: How AI chatbots restructure epistemic identity and virtue [PDF]. (2025). Poligrafi. https://ojs.zrs-kp.si/index.php/poligrafi/article/download/510/531/5537

AI and the wisdom of uncertainty. (2026, July). Psychology Today. https://www.psychologytoday.com/us/blog/urban-survival/202607/ai-and-the-wisdom-of-uncertainty

Epistemic considerations when AI answers questions [PDF]. (2023). arXiv. https://arxiv.org/pdf/2304.14352.pdf

The AI-mediated communication dilemma: Epistemic trust, social... [PDF]. (2025). PhilPapers. https://philpapers.org/archive/SAHTAC.pdf

Small talk: Is it being replaced by search-first social behaviour? (2026, January). Digital Journal. https://www.digitaljournal.com/social-media/small-talk-is-it-being-replaced-by-search-first-social-behaviour/article

Over-reliance on chatbots can diminish critical-thinking skills, study finds. (2026, June 19). The Guardian. https://www.theguardian.com/us-news/2026/jun/19/chatbots-critical-thinking-skills

Argumentative conversational agents for online discussions. (n.d.). PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC8143987/

Smartphones in relational conflicts: Examining when and how they interfere. (2024). Journal of Social and Personal Relationships (or relevant journal). https://www.tandfonline.com/doi/full/10.1080/1041794X.2024.2433507

Dealing with ambiguity in tech projects: A review of tolerance for... (n.d.). Journal of Applied Research in Management and Business. https://www.armanprocessjournal.ir/article_735646_en.html

Tolerance of uncertainty: Conceptual analysis, integrative... (n.d.). PubMed. https://pubmed.ncbi.nlm.nih.gov/28324792/

------------

As always, I'm available to help you get subscribed to Legalshield or IDshield (wracton@gmail.com)

My app, AI-ROBICS (or something like that), will bring together both the legal and web-based protection of those two with recommended systems for dealing with both the cognitive and relational ramifications of AI and its application in so many aspects of our lives today.

www. williamacton.legalshieldassociate.com

Initial prompt (with added subsequent prompts for clarification)

We earlier did a phenomenologically grounded piece that worked well. https://hipoeces.blogspot.com/2026/05/the-quiet-erasure-phenomenological.html. Using the same premise and general approach, create an analogous piece based off the "interpersonal" research we did earlier, where the voice narrating the story is a mother of three, ages 5 to 15 who has apparently become dangerously dependent on her AI chatbot to the extent that very much aware of how it is compromising virtually all of her interpersonal relationships, expecially with her family members and friends . . .and dog.

Plotting (and reversing) personal AI‑induced decline in critical thinking ability

 

Caveat emptier: This post was drafted with help from an AI assistant (Perplexity)— but ideated and edited extensively by the human, Bill Acton. The links in the text connect to research or sources that support the specific claim being made, most of which were initially surveyed by Perplexity and then confirmed by me. 

Clker.com

A Quantitative and qualitative checkup of AI use and its potential impact . . . 

You probably can’t run a brain scan, but you can watch for specific patterns in your behavior, work, and feelings that research is starting to link to AI‑related declines in critical thinking.

1. Know the risk pattern: cognitive offloading

Studies are converging on the idea that the main risk is not “using AI” but over-offloading thinking to it.  Frequent heavy AI users in several studies showed lower scores on critical thinking tests, with “cognitive offloading” letting the tool do the reasoning work, acting as the mediator.  Warning signs:   

- You routinely ask AI for explanations, arguments, or outlines before trying to reason them out yourself. 

- You notice you are skipping the cognitive struggle phase in hard tasks (wrestling with ideas, drafting, revising) and jumping straight to polished answers. 

- You increasingly feel that without AI you’d be “stuck” on tasks you once did alone. 

If those describe some or most of your serious thinking work, your critical engagement may be eroding rather than being extended or enhanced. 

2. Behavioral checks: How you work without AI

A practical way to test whether AI is degrading your critical thinking is to temporarily remove it from small parts of your workflow and observe the difference. Try experiments like:

- No‑AI sprints: Pick one task you often use AI for (e.g., outlining a post, analyzing an article, planning training) and do 20–30 minutes completely without AI, then note:  

  - How hard is it to get started?  

  - Do you stall, or can you still generate lines of reasoning and questions? 

  - Independent re-creation. Let AI generate an answer to a question you care about, set it aside, and a day later write your own answer from scratch, then compare for depth and structure.

- Paper test: For logic puzzles, argument analysis, or reading a dense text, work only with pen and paper and see whether your stamina or patience has changed over time. 

If you consistently find yourself unable to start, sustain, or organize thinking without the tool, that’s strong evidence that your independent skills are atrophying. 

 3. Product checks: How your work is changing

AI‑related critical thinking decline often shows up in the *character* of your work.

Look for trends like:

- Arguments or essays feel more polished but less yours—generic phrasing, fewer original turns of thought, less surprise. 

- You’re doing less fact‑checking, corroboration, or seeking counter‑arguments, and more copy, light edit, and send.

- Your questions to AI are shallow (“write this for me”) rather than probing (“here’s my reasoning—what am I missing?”)

You can test this directly by taking a few older pieces you wrote pre‑AI and comparing them to current pieces: are there fewer explicit reasons, counterpoints, and revisions visible in your newer work?

 4. Internal checks: How you feel while and after using AI

Several researchers and commentators note that over‑reliance on AI can correlate with mental “flattening”: less effort, less reflection, and more anxiety. Ask yourself:

- After a long AI session, do I feel mentally "exercised" or somewhat numb and passive? 

- Do I at times trust AI’s answer more than my own instincts, even on things I know well? 

- Do I feel uneasy tackling complex problems, especially those I am not as intimately familiar with, without checking with AI first? 

A growing sense that you can’t think effectively without the tool, is a soft but important indicator of diminished critical confidence. 

5. Use of a Self‑Assessment Rubric (like the one I have created below!)

( It is possible to adapt existing AI literacy and critical thinking rubrics, although none generally available focus on the impact of AI use/dependence, per se.) 

For example, rate yourself (1 = often, 3 = sometimes, 5 = rarely) on each statement:

- “I accept AI outputs without checking sources, logic, or alternatives.” 

- “I start with AI’s answer before I try thinking the problem through myself.” 

- “I could still perform my core professional or creative tasks reasonably well if AI disappeared for a week.” 

Link to the (Am I losing my edge?) A crirtical thinking checklist for AI users

Ideally, one should repeat this self‑rating regularly for evidence of the need to rebalance. 

Ideally, after a close examination of the potential impact on your digital identity, you should consider subscribing to Legalshield or IDshield, as well. (Be delighted to introduce you to both at your convenience!) 

wracton@gmail.com

williamacton.legalshieldassociate.com