Peter Yang’s Five Steps for Using AI to Help Care for His Mother

My mother has survived cancer three times, and now it has returned. During the first three experiences my parents carried most of the burden. This time I want to help: I have built a free AI skill called F Cancer to support the practical work of caregiving.

I outline five steps: make a one-page brief, organize records and scans, research treatment options, find experts who can help, and protect time with family. I see caring for a loved one as one of the most important uses for AI.

I expected to use AI mostly to research the disease. Instead, it has been particularly useful for keeping track of paperwork and follow-ups across doctors and departments. The skill maintains a concise, factual brief in a shared Google Doc.

The brief has five parts: patient information and medical contacts; the three most important next actions or questions for the doctor; what is known about diagnosis and treatment; plain-language explanations of medical terms; and a dated care log. It can be updated as new information arrives.

In a demonstration using a hypothetical patient, I show how the skill creates the brief from a description of the diagnosis and care team. The document keeps contacts, possible second-opinion information, markers and treatment status together, making it easier to prepare for short appointments.

Once the document exists, I can ask what three questions matter most for an upcoming appointment. It does not have to contain everything at the start; the goal is to keep a shared account of what we know as diagnosis and treatment proceed.

The brief depends on the records behind it. My mother’s earlier cancer history is scattered across hospitals and portals, and I wish we had requested three kinds of material sooner: written reports and treatment history; imaging files; and pathology blocks or slides from biopsies.

In my experience imaging may arrive as DICOM files on a drive or CD and can take weeks. Pathology material can take even longer, especially when an older sample must be retrieved. My mother is still waiting for a block from a previous cancer, so I recommend starting by asking the relevant institutions early.

I keep my mother’s documents and scans in a shared Google Drive. Giving AI access to the brief and records helps ground its responses in her history. It can also help make a list of what to request, find request links and draft correspondence, subject to checking the details.

My mother’s process began with a suspicious finding on a scan, followed by more imaging and a biopsy confirming cancer. A tumor board brought specialists together to discuss the case and options. Standard care commonly follows clinical guidelines and may involve different combinations of treatment.

Two months after the first scan, my mother was still waiting to learn whether this was a recurrence of breast cancer or something else. I respect the doctors, but coordinating through calls, voicemail, faxes and physical discs has been frustrating. The gap between headlines about AI and the everyday work of obtaining care is large.

While waiting, I want to understand standard care and seek a second opinion. I suggest asking AI to locate public guidance for a case and using that information to prepare for conversations with clinicians. In the demonstration, the treatment examples concern a hypothetical patient with specific markers.

I mention several major cancer centers as possible sources of second opinions. Obtaining one can itself take weeks, and organized records and images help. For my mother’s ongoing second-opinion process, I use AI to organize records and draft emails and a cover letter explaining her case.

I also discuss tumor sequencing, which looks for markers that may inform treatment decisions. Our family is considering a service recommended to us, at an estimated cost of about $4,000. I hope the resulting report will help inform later discussions; the diagnosis and treatment decisions remain matters for her care team.

Clinical trials may offer access to approaches still being studied, but I caution that early-stage trials are unproven and can carry risks. I demonstrate asking AI to search for relevant options and point to ClinicalTrials.gov. Location and the patient’s particular disease matter when evaluating a trial.

Sharing my mother’s story brought offers of help. People introduced us to others who could help navigate testing and pathology review, suggested questions, and offered clinical perspectives. I use AI to draft concise updates to the people supporting us.

Many families keep a cancer diagnosis private. Sharing as much as feels comfortable, even just with friends, can give others a chance to offer support; many people have been through something similar. Each family can decide what it wishes to share.

When I learned about my mother’s cancer, I spent a week with her, taking walks, sharing meals and talking. I call my parents every day so they can see and talk with my children. Those moments matter alongside all the administrative work.

In my early forties, with young children and aging parents, it is easy to focus on career, money and titles. This experience has reminded me that time with a spouse, children and parents is limited and easy to postpone. I want to be more deliberate about protecting it.