I helped draft an anti-AI open letter after interviewing an illustrator who was panicked about the looming extinction of her profession. It was a plea to newsrooms to respect the intrinsic value of art made by humans and the sensitivity that comes with it. It received close to 4,000 signatures and was covered in newspapers around the world. I spent the next year finding ways to fight the technology, producing podcasts and a symposium at the Rhode Island School of Design that explored the impact AI would have on art. I thought any good from this tech was illusory and not worth the downstream costs — until my father got sick.

In January, he felt pain in his back and stomach. He was taken to every doctor we could think of, then to whoever that doctor could think of. No one fixed it. If he ate, it was only dinner. At night, my mother watched him staring out from bed onto their quiet San Fernando Valley street, his back throbbing. He'd get up, pace for hours; down to the living room, into my childhood bedroom, upstairs, downstairs. By morning, we would find him upright, sleeping on the living room couch, cell phone still bright from hours of scrolling.

His doctors didn't know what it was. And no one on his care team was responding fast enough to his list of ailments. Months passed. His sturdy frame turned fragile and bony. Tests were ordered. Weeks went by with no results.

His pain left my mother speechless. She insisted I write to his doctors and threaten to cause problems if they didn't help. This request wasn't unusual; I've been put to work by my family ever since my entrepreneurial father sent my mother, eight months pregnant with me, on door-to-door sales calls, figuring her belly would move merchandise. But I didn't know where to begin; his care had become "humanless." Emails went unanswered, calls answered by bots. Against every instinct, I prompted ChatGPT for help.

I typed in a crude concoction of fragments of my mom's experience: "I go back and forth between his gastro, heart doctor, and family doctor, then they send me to his psychiatrist … the only way we are going to find out what is happening is if everyone talks to each other." Then I put in phrases I thought could add urgency: "A five-alarm fire. We need a plan!" I told it to draft a complaint that would trigger someone internally to advocate for us. Within seconds, I had a step-by-step roadmap to get a patient advocate, plus tips on how to get a quick response.

But it was too late. Within a day, he was in the ER with a rock-hard stomach, unable to keep food down.

Test results trickled in, but we didn't have enough face time with doctors to break down what they meant; the doctors were overextended and could only see us for a few minutes at a time. I knew the risks of entering personal data and pathology reports into ChatGPT, but I was becoming addicted to seeing not only instant results but also responses that generated lists of detailed questions for the rare moments when we saw a doctor in person. With AI, I wasn't helpless.

Day three in the hospital, I asked ChatGPT to interpret a new bit of information I found in his portal: "What could a small nodule in the pancreas — about 1.6 cm — be?" The response was prescient. Of the long list of possibilities, like pseudocyst, lipoma, and neuroendocrine tumor, some were potentially malignant, including pancreatic adenocarcinoma and pancreatic neuroendocrine tumor.

Early the next morning, we received a call from his doctor: It was pancreatic cancer. He was given six months to a year if he did chemotherapy to shrink the tumor. There were still things he wanted to do — visit his childhood friends one last time, see my daughter back East, spend time in his yard staring at the cacti, feeding his fish, rubbing his fingers on the wild lavender — so he agreed.

With the doctors' visits erratic and short, we used ChatGPT to strategize his care. I fed it the names of other hospitals and asked how it would tackle my father's prognosis. One on the East Coast, known for its aggressive and exacting approach, helped a family member beat breast cancer. We took its list of recommendations and used it to compare the care he was getting — and demand changes whenever there was a lapse. Meanwhile, my sister and I took 24-hour shifts to ensure we were there to advocate for him when they came into his room. His sleep worsened, and he became disoriented, unrecognizable to himself.

One evening, I stood by him as he stared at himself in the aluminum-edged mirror, tracing his temples and jaw with his index finger like he was seeing them for the first time. He asked, "Have you ever seen me this skinny?"

Were the doctors really looking at him? Could they see how weak he'd become? Or the bruises on his rail-thin arms and hands from every collapsed vein and new IV? Would he actually be able to tolerate chemotherapy?

The day he was discharged, I flew home to see my daughter. The plan was to return three days later to take him to chemo. That night, surrounded by family, he assumed his position at the head of the dining table for a dinner he barely touched. Still, we trusted what we were told: home would be a tonic, and he would begin to recover in time for treatment.

Within hours, my family knew his doctors were wrong. No medication was working, he vomited, he didn't sleep, he was in so much pain he couldn't stop moving. Panicked by his decline, my mother called me. She put me on speakerphone; I asked if he wanted to go back to the hospital. His response was incoherent. My mother chased him up the stairs holding his IV bag. He lay down on my childhood bed and looked at a shelf stacked with family photos. He told my mother to pull one down. She handed him a photograph of the two of us, five-year-old me, my cheek in profile as he kissed me. He pulled it to his chest, turned on his side and took his last breath.

My mother tells me I wouldn't have wanted to see him in his final moments. She said it was easier for him to let go without me in the room. But I'll never stop wondering: What if?

What if he got the human attention he desperately needed? What if AI hadn't been the one carefully reviewing all his tests and reports, generating endless lists telling me what to ask and expect next? What if I didn't have to rely on a computer to simulate compassion while he sat unattended by doctors for hours, crippled by his pain?

What if we'd been told it was a matter of days, not months?

Maybe then, in the end, he wouldn't have been looking at a photograph … but instead at me.