Midhat Tilawat & Farida Khalaf: Author Interview #21, AI Everywhere, Volume 1
AI Everywhere, Vol.1, Ch.8: AI in Decision-Making: A Helper, Not a Hand to Hold Forever. How AI provides vital decision support during life’s most vulnerable transitions, from postpartum to elder care
Welcome back to the She Writes AI interview series, where we talk with the authors behind AI Everywhere, Volume 1 exploring their personal journeys and professional expertise to understand how we can build a future for AI that is truly ethical, responsible, and human-centered.
This is the 21st episode in our series, and today I am joined by AllAboutAI by Midhat Tilawat. Representing the Home section of our book, her Chapter 8: AI in Decision-Making addresses the role of technology in our most private and high-stakes moments. Midhat explores how generative AI can serve as a “thinking partner” during periods of intense emotional and cognitive depletion, specifically focusing on the stressful transitions of postpartum recovery and caregiving for an ailing parent.
AllAboutAI by Midhat Tilawat brings a deeply compassionate and grounded lens to this conversation. She challenges the idea that using AI for personal decisions is a sign of weakness or a loss of intuition. Instead, she reframes it as a necessary support system that can help bridge the gap when the “human battery” is low. Her work moves us past the fear of dependency and into the realm of strategic delegation, where AI handles the data-heavy “what” so that we can remain present for the emotionally-heavy “who”.
In this conversation, Midhat explores the defining line between a helpful tool and a permanent crutch, how to use AI to navigate “information paralysis” during a crisis, and the introduction of the “H.E.A.R.T. Check” — a framework for maintaining human agency in AI-assisted decisions. Whether you are a new parent, a caregiver, or simply someone looking for a disciplined way to manage life’s complex choices, Midhat’s insights offer a roadmap for using AI as a helper while keeping your own hand firmly on the wheel.
Below is AllAboutAI by Midhat Tilawat’s full interview response to Farida Khalaf’s questions on AI in Decision-Making: A Helper, Not a Hand to Hold Forever from AI Everywhere, Volume 1,
Topic: AI in Decision-Making (Chapter 8)
Intended Audience: For parents, caregivers, and individuals navigating high-stress life transitions who seek a disciplined framework to use AI for decision support without sacrificing their personal agency or intuitive judgment.
Part 1: The Professional Path & The Collective Mission
The Expert Perspective: AllAboutAI by Midhat Tilawat, you’ve explored the role of AI in some of the most intimate and high-stakes areas of life: postpartum and elderly care. What was the “spark” that made you realize AI could be a vital helper, rather than just a technical tool, during these periods of vulnerability?
Midhat: The “spark” didn’t come from a research paper or a professional interest in AI. It came from a moment of exhaustion and emotional overload during postpartum.
I remember sitting with my newborn after hours of crying, feeling completely depleted, and realizing that I didn’t have the emotional space to process what I was feeling, let alone explain it in a way that would be understood without judgment or dismissal.
In Pakistan, mental health is still deeply taboo, especially when it comes to postpartum depression. In many Desi households, the idea itself is either minimized or not even recognized as something real. The response is often layered with cultural expectations, gratitude, resilience, sacrifice, rather than emotional validation. So instead of turning inward or outward to people who might normalize or dismiss what I was feeling, I turned to AI.
What surprised me wasn’t just the response, it was the absence of judgment. It didn’t tell me to be more grateful, it didn’t reduce my experience to cultural expectations, and it didn’t try to correct my emotions. It simply helped me put language to something I was struggling to even name.
That was the shift for me.
I realized AI wasn’t replacing human support, it was creating a rare kind of psychological breathing space in a moment where I didn’t have one. Especially in postpartum, and later in elder care, where decisions are constant, emotional, and socially loaded, AI’s role became clear: not as an authority, but as a stabilizing layer that helps you think when everything around you feels overwhelming.
That was the spark, the moment I understood AI is most powerful not when it decides for you, but when it helps you stay grounded enough to decide for yourself.
The “Home” Perspective: Your chapter is situated in the “Home” section and is titled “A Helper, Not a Hand to Hold Forever.” Why is it critical for families to understand that while AI can support decision-making, it cannot (and should not) replace the fundamental human intuition required in caregiving?
Midhat: It’s critical because caregiving decisions are never purely informational, they are deeply relational, emotional, and contextual. They are shaped by history, family dynamics, unspoken tensions, cultural expectations, and the kind of love that doesn’t translate neatly into data.
AI can process information quickly. It can summarize medical literature, list options, and even simulate empathy in language. But what it cannot access is the lived reality inside a home, the subtle shifts in a person’s behavior that only a daughter notices, the emotional cost of a decision within a joint family system, or the long-term relational consequences of a choice that may look “optimal” on paper but feels deeply wrong in practice.
In caregiving, whether postpartum or elder care, intuition is not the opposite of intelligence. It is intelligence that has been shaped by proximity, responsibility, and lived emotional context. That kind of understanding cannot be outsourced without loss.
In many families, especially in our context, there is also a growing temptation to treat AI as an authority because it sounds structured and confident. But confidence is not the same as correctness, and clarity is not the same as truth. If we start replacing human judgment with algorithmic suggestions, we risk weakening the very capacity that caregiving depends on: the ability to sit with uncertainty and still choose with care.
That’s why I frame AI as “a helper, not a hand to hold forever.” It should reduce friction, organize information, and expand perspective, but the final responsibility, the moral weight, and the emotional intelligence required in caregiving must remain human. Because only humans carry consequences, relationships, and love in a way that makes those decisions meaningful.
Part 2: Navigating Crisis with Decision Support
Reducing Cognitive Load: You highlight how AI provided decision support during postpartum, a time of extreme physical and emotional exhaustion. How can AI help filter the “noise” for a new parent who is overwhelmed by conflicting advice and sleep deprivation?,
Midhat: In those early postpartum weeks, the biggest challenge isn’t a lack of information — it’s too much of it, coming at you when you’re physically exhausted, emotionally sensitive, and cognitively overloaded.
Every person around you has advice. Every forum has a different rule. Even simple things like feeding, sleeping, or interpreting a baby’s cry come with conflicting opinions that can quickly turn into anxiety rather than clarity.
This is where AI can play a very specific and limited but valuable role: it can act as a filter, not an authority. Instead of adding more noise, it can help structure it.
For example, it can take scattered advice and organize it into categories: what is medically supported, what is cultural practice, what is anecdotal, and what is simply opinion. It can summarize long explanations into digestible points when sleep deprivation makes even basic reading difficult. It can also help you compare options side by side, so you’re not holding ten contradictory voices in your head at once.
But more importantly, it can reduce the cognitive load of processing, not the responsibility of deciding. Because when you are running on broken sleep and emotional depletion, even small decisions start to feel overwhelming. AI can create a pause, a structured space where chaos becomes readable again.
However, the key boundary remains essential: AI should not become the voice that replaces judgment. It should be the one that quiets the noise enough so your own intuition, which is still the most context-aware system you have, can come back online.
In that sense, its role is not to tell a new parent what to do, but to help them hear themselves again amidst all the noise.
The Caregiving Complex: In your experience caregiving for an ailing parent, what were the most stressful “decision gaps” where AI provided clarity that traditional resources might have missed?
Midhat: In caregiving, I didn’t experience decisions as moments, I experienced them as a continuous background noise I was always trying to make sense of.
The most stressful parts were the “decision gaps”: small but constant uncertainties where there was no clear answer, only fragmented information, emotional pressure, and limited time to think. AI helped most in those gaps.
One was information fragmentation. Medical notes, prescriptions, and family observations were all scattered. AI helped me bring them into one structure, timelines and summaries that made patterns easier to see when my own mental bandwidth was low.
Another was translation, not language, but meaning. It helped turn clinical medical explanations into practical understanding of what I should actually look for at home and what required urgent attention.
The third was communication within the family. In a joint family system, caregiving discussions can become emotionally charged. AI helped me frame messages more clearly and neutrally so conversations became less reactive and more structured.
But even then, it stayed a tool for clarity, not truth. Everything still had to be grounded in lived reality and medical judgment. Because in caregiving, AI can reduce noise, but it cannot replace responsibility for what is real.
Information Synthesis: When navigating a parent’s health, information can be fragmented across doctors and records. How did you use AI to synthesize this data into a manageable “roadmap” for care?
Midhat: In caregiving, I stopped thinking in terms of “information” and started thinking in terms of “trajectory.”
Everything was coming in fragments, reports, prescriptions, observations, but what I needed wasn’t more detail. I needed to understand what was changing over time.
So I used AI to convert scattered inputs into a simple progression view: what changed, when it changed, and what might be connected. Instead of re-reading every document from scratch, I could see the story of the illness unfolding in one place.
The real value wasn’t summarization, it was continuity. It helped me stop restarting my thinking every time a new piece of information arrived.
That shift quietly changed how I made decisions. I was no longer reacting to isolated updates; I was responding to a developing pattern.
But I still treated it as a working map, not a final truth. The “roadmap” only made sense when I constantly checked it against real behavior at home and medical feedback. Without that grounding, structure can look like clarity even when it isn’t.
The goal isn’t just to get an answer; it’s to free up the mental space required to be fully present for the people who need us.
Part 3: The Boundaries of Artificial Support
Helper vs. Replacement: You are clear that AI is not a “hand to hold forever.” What are the specific “red lines” where a caregiver must ignore an AI suggestion and rely solely on their “human signal” and lived experience?
Midhat: For me, the red lines are very clear: any situation where lived reality contradicts the model.
If AI suggests something that doesn’t match what I am seeing in the room, behavior, pain, distress, confusion, I don’t debate it. I drop it. Caregiving is one of those domains where the body in front of you is always more authoritative than any output on a screen.
The second red line is emotional safety. If a suggestion increases panic, guilt, or paralysis instead of clarity, it is not useful, even if it is “technically correct.” In caregiving, a decision that cannot be executed calmly is already the wrong decision.
And the third is context mismatch, especially cultural and family realities. If advice ignores feasibility within a real household (time, finances, family dynamics), it is not actionable knowledge; it is abstraction. And abstraction cannot be the basis of care.
So my rule became simple: AI can inform, but it cannot override lived observation, emotional stability, or real-world constraints.
The Empathy Gap: AI can provide “accurate” data, but it cannot feel the weight of a caregiver’s struggle. How do you prevent yourself from becoming “de-personalized” when relying on algorithms for life-and-death care decisions?
Midhat: The risk isn’t that AI replaces caregivers, it’s that it slowly flattens the emotional weight of what you’re carrying.
When you are tired, there is a temptation to treat structured answers as emotional relief. But caregiving is not just problem-solving; it is grief, responsibility, fear, and love happening at the same time.
I prevent depersonalization by doing something very intentional: I always bring AI output back into the human world before accepting it. I don’t just ask “Is this correct?” , I ask, “Does this still feel true in the room I am standing in?”
Also, I never let AI become the final emotional interpreter. It can help me organize thoughts, but I don’t let it name my feelings for me. Because the moment you outsource emotional language, you also start outsourcing emotional ownership.
In short: AI can structure the situation, but I stay responsible for the meaning of it.
Managing “Predictive Feeds”: If a caregiver is already stressed, a predictive feed might show them worst-case scenarios. How do you use AI as a finite tool to avoid the “social media spiral” that others, like Brie-Anna Willey, warn about?
Midhat: There is a well-documented concept in research called cognitive overload, where the volume and complexity of information exceeds our ability to process it in a meaningful way. In AI systems, this often shows up as what researchers describe as information expansion, where tools don’t just answer a question; they multiply possible outcomes and scenarios, increasing anxiety rather than clarity.
A recent paper on cognitive overload in the age of AI even notes that modern systems can unintentionally amplify anxiety by increasing “complexity, information proliferation, and perceived lack of control.”
That idea feels very real in caregiving.
When I was already stressed, especially in a Pakistani household where there is constant noise of opinions, advice, and expectations, AI could easily become another layer of that overload. A simple question about a symptom or medication could quickly expand into worst-case scenarios, rare complications, and long lists of “what could go wrong.”
So I started treating AI as a finite tool, not a predictive feed.
Practically, that meant I stopped engaging with it for open-ended exploration. I would not ask “what are all possible outcomes?” Instead, I would ask, “what are the 2–3 things I need to pay attention to right now?” That shift kept the output usable instead of overwhelming.
There is also research in human-AI interaction showing that when systems increase cognitive load, they don’t necessarily improve decision quality; they can actually reduce it, because users either over-trust or disengage completely under pressure.
In caregiving, that’s exactly the danger: when you are already emotionally depleted, too much information doesn’t empower you — it destabilizes you.
So my rule became very simple:
AI is allowed to clarify, structure, and narrow, but not expand endlessly.
And I always close it after I get what I need. Because in caregiving, especially in high-pressure home environments, the goal is not to simulate every possible future. The goal is to stay steady enough to act in this one.
Part 4: Governance, Ethics & The Human Signal
Data Privacy in Care: Caregiving involves sharing incredibly sensitive health data with LLMs. What is your advice for parents and caregivers on maintaining governance over their family’s digital privacy?
Midhat: In caregiving, especially during postpartum or elder care, people don’t think in terms of “data privacy”, they think in terms of urgency. You are tired, overwhelmed, and trying to make sense of medical information quickly. In that state, it becomes very easy to paste prescriptions, reports, or detailed health histories into AI tools without thinking about long-term privacy implications.
My advice is simple: assume permanence.
Anything you share with a system should be treated as if it could exist beyond the immediate moment of use. That mindset alone changes behavior.
Practically, caregivers don’t need to avoid AI, but they do need boundaries. For example, removing identifying details, avoiding full medical records when summaries are enough, and using AI for structure rather than raw sensitive data whenever possible.
But the deeper governance issue is this: families should decide in advance what kind of information is appropriate to externalize and what must remain within human trust networks, doctors, immediate family, and official medical channels.
Because once sensitive care data becomes habitual input into digital systems, it stops feeling sensitive, and that is where the real risk begins.
Confronting Bias in Health: We know AI reflects the biases of its creators. How might a caregiver encounter “cultural absence” or medical bias when seeking advice for a new mother or an elderly parent from a marginalized background?
Midhat: One thing caregivers quickly discover is that AI systems often respond from a “default world”, and that default world is not ours.
For example, advice around elder care may assume access to nursing homes, respite care services, or professional home support. Postpartum guidance may assume nuclear family structures, independent mobility, or mental health systems that are accessible and stigma-free. In many South Asian or Pakistani households, these assumptions simply don’t hold.
This creates what I would call cultural absence bias, not just incorrect information, but missing context.
So instead of asking, “Is this advice correct?” a caregiver has to ask, “Is this advice even designed for a situation like mine?”
In my own experience, AI would sometimes suggest solutions that were technically sound but socially or practically unworkable — not because the model was wrong, but because it was operating without awareness of household dynamics, financial constraints, or cultural expectations like joint-family decision-making and “log kya kahenge.”
This is where caregivers need to actively reinsert context that AI does not naturally carry. Otherwise, you end up with advice that looks universal but is actually detached from your reality.
Relational Responsibility: Sydnor Hain (they/them) (Chapter 3) argues that relational ethics matter more than rules. How does your experience with AI support the idea that the “human in the loop” is the only one who can carry the ultimate ethical responsibility for a decision?
Midhat: My experience strongly supports the idea that relational ethics cannot be outsourced and AI makes this even more visible.
In caregiving decisions, especially for a parent with Alzheimer’s or a newborn mother under stress, there is always a point where the decision stops being informational and becomes relational. It is no longer about what is “optimal” in theory, but what preserves dignity, trust, and long-term emotional consequences within a family.
AI can assist in mapping options, risks, and structures, but it cannot participate in the relationship where the consequences will unfold. It does not sit in the room. It does not face the person. It does not live with the aftermath.
That is why the “human in the loop” is not just a technical requirement, it is an ethical necessity.
Because in real caregiving, responsibility is not about having the best answer. It is about being the one who will still be there after the answer is chosen, to carry its emotional and relational weight.
AI can inform decisions, but only humans can be accountable to them, to the patient, to the family, and ultimately to their own conscience.
Part 5: Interactive Closing & Actionable Awareness
The Final Message: If a reader is currently in the “thick of it” struggling with a new baby or a sick parent, what is the one way you hope they use AI this week to reclaim a bit of their own mental space and agency?
Midhat: If I could leave one thing with someone who is in the thick of it, a new mother, or someone caring for a sick parent, it wouldn’t be a big framework or a theory.
It would be this: don’t use AI to carry your life. Use it to put it down for a moment.
I remember those days as a kind of mental noise that never stopped. Even when the house was quiet, my mind wasn’t. I was always remembering something, a medication, a feeding detail, a comment someone made, a worry I didn’t have time to fully think through. It felt like I was constantly holding ten invisible things at once.
In that state, AI helped me most in a very simple way: I would dump the mess into it. Not for answers. Just so I didn’t have to hold everything in my head at the same time.
And what came back wasn’t always “truth”, but sometimes it was clarity. A way to see the situation in front of me without the emotional fog.
But I also learned something important: clarity is only useful if it gives you your agency back, not if it replaces it.
That’s why I started using what I call the H.E.A.R.T. Check, not as a system, but as a pause I force myself to take before I follow any AI suggestion:
H — Human goal first — What do I actually want right now?
E — Evidence request — Is this evidence or just confidence in language?
A — Alternatives and adversarial view — What is the other side of this advice?
R — Reality filter — Does this even fit my real life, my family, my limits?
T — Take responsibility — Am I still the one deciding this, or have I handed it over?
Most days, I don’t even “complete” it formally. It’s just a breath I take before I move from thinking to acting.
Because the real danger in caregiving is not lack of information. It’s losing yourself inside too much of it.
And if AI can give you anything in that moment, I hope it’s not more noise. I hope it’s just enough space for you to feel like you again, even for a few minutes.
AI is a helper, not a hand to hold forever. It provides the scaffolding while we find our footing again. - Midhat Tilawat
Why This Matters
AllAboutAI by Midhat Tilawat’s perspective is a vital reminder that the “human signal” in decision-making is responsibility, not just calculation.
Decision Support in Crisis: Midhat identifies AI as a critical intervention for “cognitive overload.” By using tools to organize medical information or postpartum schedules, caregivers can reduce the “ecosystem tax” of hidden labor that leads to burnout.
A Helper, Not a Crutch: She emphasizes that while AI can provide the “support,” it must not become a “hand to hold forever.” The goal is to use the technology to move through a period of vulnerability back toward sovereignty and self-reliance.
The H.E.A.R.T. Check: This original framework ensures that AI-assisted decisions remain aligned with personal values and ethics. It serves as a mental audit to verify that the “human signal” remains the final authority before an action is taken.
Healing Information Paralysis: In high-stress scenarios like elder care, the abundance of conflicting information can be paralyzing. Midhat shows how AI can synthesize complex data into actionable insights, allowing for faster, more confident human choices
Thank you, AllAboutAI by Midhat Tilawat, for your vulnerable and deeply practical contribution to the Home section of our book. For readers who have ever felt “drowned” by the demands of caregiving or new parenthood, her chapter in AI Everywhere, Volume 1, is a beautiful guide on how to reclaim your agency through purposeful technology use.
Don’t miss her ongoing work and her newsletter, All About AI, where she writes about AI as it really shows up—in workflows, in search, and in everyday life. Her final lesson is one we should all carry: AI can give you the data, but it cannot give you the “why”: that remains your unique human superpower. I personally highly recommend her H.E.A.R.T. Check approach as a way to turn AI from a cold algorithm into a warm, supportive partner for the home.
And if you’re interested in more conversations like this, consider subscribing to She Writes AI. I’ll continue sharing insights from the voices shaping the future of AI, including contributors from AI Everywhere, Volume 1, where 26 women from across five continents explore how artificial intelligence is changing the world we live in, and how we can change it for the better.
References
Book purchasing information
Buy the book (now on Amazon, Gumroad, Lulu, Ingram Spark, and many ebook platforms). Look for our upcoming abridged audiobook edition in July!
Tips from our publisher on how to buy an Amazon ebook from a non-KDP-marketplace country, and how to read it without a Kindle device (free)
Review links (we’re on Amazon, Goodreads, and Storygraph)
For more insights on the book, see aiEverywhereBooks.com - our beautiful book series website created by Blessing Okpala, PhD:
Thanks for reading She Writes AI Community! Subscribe (free) to receive new interviews, weekly digests, & book news and to support our work:
 | A guest post by
|