You’ve probably already used AI today without thinking about it. A search result that seemed to read your mind, a suggested reply that sounded like you, a recommendation that landed a little too well. These systems are woven into daily life in ways that are easy to overlook precisely because they’re designed to be. I’ve spent the better part of a decade building AI systems for professionals, and that work has shown me both sides: genuine capability that I think is worth celebrating and emerging risks that deserve more attention than they’re getting. Here are a few field notes from the inside about what’s genuinely worth being excited about and what I think we should be watching closely.
AI is Turning Curiosity Into Capability
A first-generation college student can now get tutoring that rivals expensive prep schools. A small business owner can draft a contract or troubleshoot a supply chain problem without a retainer. A patient can walk into a doctor’s appointment with better questions because they understood the research beforehand. The access gap isn’t gone, but it’s narrower than it was — and that’s worth defending even as we push for guardrails.
But Inference is More Dangerous Than Disclosure
Most people worry about what they’ve directly shared online. The scarier reality is what AI systems can deduce without you ever saying it. Typing cadence, scroll behavior, and the time of day you engage with certain content can reveal mental health status, financial stress, and political “persuade-ability” with surprising accuracy. That inferred profile carries the same commercial value as anything you’ve consciously shared, often without the same legal protections.
Personalization and Manipulation are the Same Mechanism
The technology that makes a recommendation feel perfectly tailored to you is identical to what makes a predatory financial offer or a misleading political ad effective. There is no technical distinction between helpful and exploitative. It’s the same model, optimizing for engagement, pointed in different directions by whoever controls it. The framing of information itself, what you see, in what order, and with what emotional weight, is being continuously optimized around a behavioral profile built over years.
And We are Losing Our Grip on Shared Reality
When synthetic media becomes indistinguishable from real media, the rational response is to trust nothing, which may be more destabilizing than any single piece of misinformation. Institutions that depend on shared evidentiary standards, including courts, journalism, and elections, are not prepared for an environment where video, audio, and text can all be plausibly disputed.
So, What Do We Do With This?
None of this means the technology is inherently bad — in fact, my opening statement exists because I genuinely believe it isn’t. But informed communities have helped steer many technology transitions in modern history, and this one is no different. Stay curious about what these tools can do for you. Stay skeptical of what you’re shown and who benefits from showing it to you. And pay attention to AI legislation while public input still matters. The neighbors who understand this technology well enough to ask hard questions about it will be the ones who help shape what it becomes.
Tom Cuppernull has lived in Beaumont-Wilshire with his partner, Katie, since 2021; he’s been on the BWNA board since 2024. Morning walks to Prince Coffee, picking up games at Puddletown, and taking his pup Oakley to Wilshire Park are some of his favorite neighborhood activities.