AI can give us more information, analysis, and advice than ever before. Why hasn’t that made hard decisions easy?

Estimated reading time: 9 minutes

By: Dr. Jamie Long

When my nearly 15-year-old Labrador, Frankie, became suddenly sick, I did what any perfectly calm and rational dog mom in 2026 would do:

I created a ChatGPT project called Frankie.

She had vomited four times. She had diarrhea. She was profoundly weak and, at one point, could barely stand. 

I reported every new symptom and observation as though ChatGPT had joined Frankie’s medical team.

I asked which symptoms could be monitored and which ones meant she needed emergency care. When her condition crossed that line, ChatGPT encouraged me to take her to the ER.

It could not examine Frankie or replace a veterinarian. What it could do was help me organize what I was seeing, recognize the red flags, and know which questions to ask when fear was making it hard to think.

During the first $1800 ER visit, the X-rays showed something we had not expected: a large mass in her chest, suspicious for a primary lung tumor.

I uploaded the records and kept asking questions. How concerning was the mass? What could a CT scan tell us? Would it require anesthesia? What might a biopsy or surgery involve for a dog who was almost 15?

The information helped. It gave me language for conversations with her veterinarians. It helped me distinguish what was known from what was still uncertain. It helped me understand the options in front of us.

And still, a new decision seemed to appear every few hours.

When she remained severely lethargic the next day, do we take her back to the ER knowing another visit could cost another thousand dollars or more? Do we hospitalize her? Do we put a nearly 15-year-old dog through anesthesia, a CT scan, a biopsy, or thoracic surgery to learn more about a mass we might never aggressively treat? How much testing would help Frankie, and when would it begin helping us more than it helped her?

AI could explain each option. It could summarize the available evidence. It could help me think through possible outcomes.

But it could not know Frankie the way my husband and I did. It could not decide what quality of life meant for her or which burdens felt acceptable to us. And it could not make the uncertainty, responsibility, or possibility of regret disappear.

At some point, we still had to choose.

AI helped me make more informed decisions. It did not spare me from having to make them.

We have never had this much help thinking

My use of AI may have been intense, but it was hardly unusual.

According to Pew Research Center, 49% of American adults now use AI chatbots such as ChatGPT, Gemini, or Copilot. Nearly one-quarter use them every day. ChatGPT use alone more than doubled between 2023 and 2026.

We are not limiting those conversations to recipes and email drafts. Anthropic analyzed one million Claude conversations and found that about 6% involved people seeking personal guidance about what to do next. More than three-quarters of those conversations centered on health and wellness, careers, relationships, or personal finance.

Health questions are already a major use case. A West Health and Gallup survey found that one in four American adults had used an AI tool or chatbot for health information or advice. Among recent users, 59% had used AI before visiting a doctor and 56% had used it afterward.

So while public conversation swings between amazement and fear about what AI may eventually do, many of us are already asking it what we should do.

Should I take this job? Is this relationship healthy? Can I afford this house? Should I be worried about this symptom? Which doctor should I choose? Was that text rude? Am I overreacting?

I understand the appeal because I feel it too. A thought that might once have floated through my mind and disappeared can now become a full consultation in seconds.

Google gave us results and AI gives us a conversation

Information gathering has evolved.

A search engine gives you pages to review. A chatbot responds to your exact wording. It remembers the context of the conversation. It answers the follow-up question, reconsiders when you add another detail, compares the options again, and can argue the other side if you ask.

It is personalized, responsive, and infinitely patient. Unlike our friends and family, it never says, “Jamie, you’ve already asked me what I think about this six times.”

That can be enormously useful. During Frankie’s illness, it helped me turn panic into specific observations and better questions.

But the conversation has no natural stopping point. If the answer does not make us feel better, we can adjust the prompt. Add more context. Ask for the strongest argument against the recommendation. Open a new chatbot and see whether it agrees. Each response can feel like progress, even when we are circling the same fear.

There’s a point where one more prompt stops being research and starts becoming reassurance-seeking.

How to tell whether you need information or reassurance

Information-seeking has a purpose. There is a material fact we do not know, it is possible to learn it, and the answer could reasonably change our decision.

Reassurance-seeking has a different purpose. We understand the basic options, but we still feel afraid, responsible, doubtful, or exposed. So we ask the question again in a slightly different way, hoping the next answer will remove the feeling.

The relief is real. It is also usually temporary. As soon as uncertainty returns, we need another answer.

Before beginning another round of research, it can help to ask:

• What specific information am I still missing?
• Could that information realistically change my choice?
• Is the information available, or am I trying to predict something that cannot be known yet?
•  Am I looking for new evidence, or am I hoping someone will make this feel safe?

Sometimes the answer will be that more research is warranted. Sometimes we need a medical opinion, financial numbers, legal guidance, or time for circumstances to change.

But sometimes information has already done its job. What remains is the emotional work of deciding.

What AI can help with and what still belongs to us

AI can make the early stages of a decision much better. It can help us:

• Find and organize relevant information
• Compare options using the same criteria
• Identify blind spots or questions we have not considered
• Separate facts from assumptions
• Prepare for a conversation with an expert 

Those are meaningful advantages. Better information can absolutely improve a decision.

But AI cannot decide which values matter most to us. It cannot determine which tradeoff we are most willing to accept. It cannot make everyone approve, guarantee that nothing painful will happen, or live with the outcome afterward.

AI can tell you what to consider. It cannot tell you what you are willing to lose.

That is especially important because the hardest decisions usually contain a real tradeoff on both sides.

Leaving may bring freedom and grief. Staying may preserve security and cost us growth. Saying no may protect our limits and disappoint someone we love. Pursuing more medical testing may provide answers and place an enormous burden on a fragile body.

No amount of intelligence can turn a genuine tradeoff into a perfect option.

A good decision is not the same as a good outcome

Thankfully, Frankie improved. Within days she was walking, eating, and looking much more like herself. I was enormously relieved.

But her recovery did not retroactively prove that every decision we made was correct. If she had deteriorated, that would not automatically have proven we chose wrong.

We had to evaluate our decisions based on what we reasonably knew at the time, Frankie’s age and medical history, the likely benefits and burdens of each option, our financial reality, and the kind of life we wanted to protect for her.

This is one of the cruelest features of decision-making: we often judge the quality of a choice using information that became available only afterward.

A thoughtful decision can produce a painful outcome. A poorly considered decision can occasionally work out beautifully. Outcomes are important, but they are not a clean report card on the quality of our thinking.

A good decision is not one that guarantees a good outcome. It is one you can understand and stand behind based on what you knew when you made it.

Information is not the same as decision-making skill

Many of us have understandably carried the belief: If I can find the right information, the correct choice will become obvious.

AI makes that belief particularly seductive because the supply of information, interpretation, and advice now feels endless.

But good decision-making asks more of us than good research. We have to identify the actual decision instead of trying to solve our entire future. We have to move without complete certainty. We have to listen to other people without turning their advice into permission. And we have to accept that choosing may bring doubt, disappointment, grief, or regret even when the decision was thoughtful.

Those are not personality traits that some people inherited and others did not. They are skills. They can be practiced systematically.

I still use AI a lot. I would use it again if Frankie became sick tonight. It is too useful to dismiss, and I do not think the answer is to romanticize the days when frightened people had fewer ways to understand what was happening.

We simply need to know which job we are asking it to do.

AI can help us think. It can help us become informed. It can even show us that we have been thinking about a problem in the wrong way.

It cannot make a consequential choice emotionally risk-free. That part of decision-making is still ours.

AI could parse every record I uploaded. It was not the one who had loved Frankie for almost 15 years. 

Find out where your decision is getting stuck

If you have researched the options, asked AI, polled the people you trust, and still cannot move forward, you may need a better way to understand what is preventing you from choosing.

I created a free guide called 4 Questions to Ask Before Making a Hard Decision. It will help you identify where you are stuck. 

The goal is not to make every decision easy. It is to get seriously good at making hard ones.

---

---