Category: technology

The Other AI: Autonomy and Influence

My son has been asking more frequently about living by himself. We’ll have a talk about independence and responsibility, and loosely talk about goals to help him move in that direction. But I also watch as he struggles to remember whether he had taken his medication, or put on deodorant, or pull his sheets up when he makes his bed.

As I watched him try to piece it together, I thought about the technology that I work with and whether it could help him.

I’ve been involved with computers and technology for most of my life, building products with bits and bytes of code and data. For the past ten years, I’ve worked in the evolving field of artificial intelligence (AI).

I recognized early on that AI could potentially transform my son’s life. As the technology matured, I watched it advance the state of medicine and healthcare.

Today, AI algorithms power diagnostic tools, accelerating the time to detect, identify, and treat complex medical conditions. AI is accelerating drug discovery, helping researchers identify promising treatments faster than ever before. It is also being used to examine genetic data to identify the right medication and dosage for individual patients.

AI could improve his quality of life in ways that weren’t possible only a few years ago. Pattern recognition can alert us when he misses a medication or a meal. Personal assistants can provide reminders, keep him on task, and communicate with him in a way that he understands. Self-driving cars will give him mobility and access to a wider world. AI-driven tools can assist him with complex tasks, help him communicate ideas, and give him greater autonomy and independence.

That’s the promise and the potential.

But here’s the problem. We live in a world where AI is already causing harm.

Inherent challenges with the technology, especially with generative AI (e.g., ChatGPT), result in hallucinations where the algorithm makes things up. The black-box nature of these algorithms makes them unpredictable and impossible to test fully, resulting in harmful behavior. And these algorithms are owned by corporations who control the data, usage, and output and can tune it to fit their agenda.

Beyond technology, people have been using these tools for nefarious purposes. It’s easy to create a false but believable story and share it on social media. It’s also easy to create completely believable but fake images and videos to mislead viewers. These bad actors are using the technology to push false narratives and generate mistrust and dissent in society.

My son struggles with memory and executive functioning. It impacts his ability to reason and determine whether what he is reading is fact or opinion, truth or lies. While I think society at large has lost its ability to thing critically, people like my son are especially susceptible to these false narratives and the harm they can cause.

So while I’m building the future with AI, I’m also guarding the present for my son. I want him to have access to all the promise this technology offers — the support, the independence, the chance to live on his own — without falling victim to its dangers. I have to be his guide, his filter, and his advocate.

Because while AI might one day help him remember his medication or build a career, it won’t teach him who to trust, what’s real, or what truly matters. It’s my job to walk beside him, protect him, and help him make sense of a world that’s changing faster than any of us can keep up with.

Yet

“Yet” is such a powerful word.

“Yet” allows us to acknowledge current struggles while leaving space for future possibilities. It’s a bridge between what is and what could be, subtly shifting focus from a fixed state to one of potential.

There have been so many times when we thought we were out of options when it came to treatments for our son’s epilepsy. We tried all the medications. We tried the ketogenic diet. Because his seizures were generalized, he wasn’t a candidate for surgeries that are available to people who have focal seizures.

Each time one of the treatments failed to control his seizures, we felt resigned to give up hope.

But even in the 10 years since my son was diagnosed, there have been many new advancements.

Genetic testing is being used to identify specific genetic mutations associated with epilepsy, which has enabled personalized treatment strategies, improving efficacy and reducing side effects.

Epidiolex was introduced in 2018 to treat Lennox-Gastaut syndrome and Dravet syndrome.

A new surgical technique called Laser Interstitial Thermal Therapy (LITT) that uses a laser to target and ablate seizure-causing brain tissue precisely was approved, reducing recovery time compared to traditional surgery.

Although it was used off-label earlier, Vagus Nerve Stimulation (VNS), which involves implanting a device that stimulates the vagus nerve to reduce seizure frequency, was FDA-approved in 2017 to treat children as young as 4 years old with drug-resistant focal epilepsy.

Deep Brain Stimulation (DBS), traditionally used to treat movement disorders like Parkinson’s, is now being used to target brain regions like the anterior nucleus of the thalamus to reduce seizure frequency in individuals with drug-resistant epilepsy.

We’ve benefited from these advancements. In addition to genetic testing, my son had VNS surgery when he was nine and DBS surgery right before his 15th birthday, even though they weren’t options when our journey started.

Of course, it’s easier to be on this side of it and say that I always had hope or that I automatically added the word “yet” to the sentence “there is nothing left to try.” I didn’t. I was overwhelmed because everything we tried didn’t stop the seizures. I had almost lost my son, and I was afraid that, with every failed treatment, every door was closing on his future.

Eventually, I would be reluctant to try a door because that would mean fewer were available. At least with untested doors, there was hope. It’s like in high school, when there was a girl I liked, if I didn’t ask her out, she couldn’t say “no,” so there was always hope for a “yes.”

What changed for me was seeing the advancements and having them offered. First, there was a new medication, then another, and then the VNS. I saw first-hand how continued progress created more doors, making trying one a little less scary. I began to believe there would be more doors, which made it easier to believe in the word “yet.”

Right now, our hand is on the DBS door. We’ve cracked it open and are waiting to see what is on the other side. There are never guarantees, but we hope it improves our son’s quality of life. If it doesn’t and we have to close that door, too, when the feeling comes back that there is nothing left to try, I will remember to complete the sentence:

It feels like there is nothing left to try…yet.

Sneaky

For a while, I’ve been thinking that Apple’s Parental Controls were poorly implemented. We would set the controls, and then within a few days or weeks, we would see them change.

As an engineer, I set out to troubleshoot the problem. At first, I thought it was because our son has two devices. Perhaps the changes weren’t propagating to one of them, and that device was corrupting the settings. But even after I ensured both devices matched, the setting would change again after a few days.

One day, my wife casually mentioned giving my son her phone because he needed to “check something.” That’s when we figured out he had convinced my wife to let him use her phone and was changing the settings.

Sneaky.

Once we figured that out, we sat him down and told him what we learned. We covered issues with trust, money (Apple’s refund policy for something like this is less than generous), responsibility, and consequences. We took away his devices for almost a week, and when he got them back, we didn’t encounter any issues.

For a while.

The other day, my son said, “Dad, I like your drawings.” I’ve been drawing on my iPad, which I leave downstairs in the living room. He knows my passcode but not the parental control code. We chatted about respecting people’s privacy and boundaries and not using my iPad without asking, and I left it at that.

A few days later, I discovered that the parental control settings had changed again. As I went in to fix all the settings, I realized that once you unlock the iPad, it didn’t require the parental code to turn off the parental controls. You can turn off the parental controls with the device’s passcode, which my son knew.

When my son saw my drawings on my iPad, he was on it to disable the parental control code. His mention of my drawings turned out to be his undoing,

Sneaky.

Controlling his impulses and urges is hard enough for him as a teenager. My son’s intellectual and emotional challenges mean that the process will take more time. We’ve been at this long enough to understand that it’s not a flip that we will switch; it’s a gradual undimming until the light is at its full brightness.

As we talked to him about these choices and the consequences, I felt conflicted. His ability to troubleshoot the problem he was facing and come up with multiple solutions is actually impressive. I know engineers with less tenacity and creativity. Using that ability to be sneaky and violate our rules, however, is the wrong expression of those skills.

Our job as parents is to help him use his power for good.