Two Realities: Living Between AI’s Promise and Its Cost

I now talk to AI more than I talk to my coworkers.

Somewhere this year, I crossed that line without noticing. I’m leading AI governance at Vanta, and AI fills my workday: the roadmap, the meetings, the tools I use to get through both.

Then I close my laptop, open my phone, and hear a different story. My community is talking about what data centers are doing to their neighborhoods, their energy bills, their water. My sister sent me an article about a recently approved data center near where I grew up in NJ. Residents are fighting back. The conversation is much closer to home than a product roadmap.

I felt that tension watching the backlash to an AI influencer retreat. I understood the criticism. But as people started shaming the influencers for attending at all, I felt torn. Some were people I followed because they made AI useful and understandable. I valued what they were teaching. I also shared some of their critics’ concerns.

That’s where I keep getting stuck. I’m frustrated that the industry I work in isn’t grappling enough with the harm. I’m also worried that people I love will get left behind while the people already ahead keep getting better at using these tools.

I want the people in my community to feel free to learn these tools and still demand better from the companies building them. And honestly, those of us in tech need to touch grass a little. 😅 How are our decisions affecting people beyond our users and our bottom line?

I don’t know how to make those two conversations meet. But the AI race is inevitable, and waiting until I feel good about all of it doesn’t feel like an option.

So I’m trying to figure out what I can do on a Tuesday. What questions to ask in the room where decisions get made. How to help the people I love learn to use this. Where to push back.

What I wish we asked instead of “Is AI good or bad?”

I keep getting pulled into some version of the same argument: is AI good or bad? My work life supplies one answer. My phone, friends and family supply another.

Neither tells me how to use AI in a way I feel good about, help the people I love benefit from it, or push back on the harm I see.

So I started with innovations I already had strong feelings about: guns, schools, cars, and nuclear weapons. I started with that range because some felt clean-cut: schools vs. nuclear weapons, while others felt more complicated like cars. Why do I want more people to have access to some and tighter limits on others? What harms am I willing to accept, and who am I expecting to live with them?

I wanted to understand my own reasoning before applying it to AI. These are the questions I landed on:

1. If everyone had access to this tool, would society get better or worse?

2. What happens to power when only a few have access or control? And who are those few?

3. What harm can it do, who bears that harm, and can it be undone?

4. Would I want my future child living in a world where this is normal?

That last one helps me think past my own lifetime.

Take cars. They help people get to work, see family, and travel to further places faster. They can also kill people. Licenses, seatbelts, and speed limits are ways we try to limit that harm. The benefits matter. So does what we ask people to risk to get them.

These questions don’t give me a pass-or-fail score. They help me see where my support comes with conditions. I can value what a tool does and still ask who’s getting rich, who’s taking the risk, and what protections we need.

You can ask these questions about a tool you use, a feature your team is building, or a data center proposed near your home. The more specific the decision, the harder it is to stop at “innovation” or “this is bad.”

I want to help more people benefit from these tools and take their potential for harm seriously. I’m frustrated that the conversations around me keep pulling those responsibilities apart.

What those questions reveal about AI

I see the case for AI making my own life better. In my job search for example, I use AI to find more opportunities that fit what I want, prep faster for interviews, and spend less time on follow-ups. Hours that used to disappear into logistics now go toward the parts of the search that need more of me.

But making a tool available doesn't mean everyone gets the same chance to benefit from it. People like me already have a head start: we know what to try, we use it daily, and we have people around us sharing what works. I worry about that advantage compounding into more income and opportunity while the people I love are still being told they should feel ashamed for using it. And learning the tool doesn't protect you from what it does to your job. You can get better at the work while your employer looks for a way to do it with fewer people. "It'll create new jobs like every other innovation" doesn't reassure me. Which jobs, for whom, and what happens to people in between? A new category of work ten years out doesn't pay someone's rent next month.

When I started writing this, job loss was my biggest worry. Since then, the possibility of AI helping someone cause catastrophic harm, including through biological weapons, has moved to the front of the discourse. My own productivity gains don't tell me whether a capability is safe to release. Some harms can't be undone, and that has to be a responsibility the people building these tools carry before the rest of us ever touch them.

My future child makes this less abstract. I want them to have tools that help them learn and create. I also want them to develop their own judgment, and a world left to live in at all.

So I come out of these questions wanting more people to benefit from AI, and more from the people building it: a community that can learn without endorsing every decision behind the tool, an industry that thinks past its users and its bottom line about who bears the risk.

I don't trust either will happen on its own.

Bill Gates put this tension plainly in his essay on the AI transition: AI could become “the greatest equalizer ever invented, or the worst source of injustice.” He argues we have no plan for the transition. His proposals are macro: tax AI tokens and robots, reserve certain jobs for humans, get the US and China in a room.

Those are decisions I want people with power working on. None of them are things I can do on a Tuesday.

But I work inside this industry, and I have a community outside it. What responsibility do I have while we’re waiting for a plan?

What I’m actually doing

I ask the uncomfortable question in the room where the decision is made.

Product decisions move fast. It’s easy to spend an entire meeting talking about what we can ship without asking who might bear the cost.

Who benefits from this? Who could it hurt? If it goes wrong, can we fix it?

I try to bring an alternative, too. A different design, a slower rollout, a way for someone to get help. I want the concern to change the decision, which means helping the room figure out what to do with it.

I teach the people I love to use it.

This is the one that gets me the most pushback, and I’ll defend it hardest.

I want to start with something that would actually help them. Something tedious in a job search. A task that keeps eating up their evening. A place where having help would give them time back.

Their concerns deserve a conversation. As do the possibilities. I want people to have enough experience with these tools to make their own choices about where they’re useful and where they don’t belong.

I pay attention locally.

The data center near where I grew up made this concrete. These decisions have an address. There are people who will live with the consequences long after the announcement.

I want to know what’s being proposed, who gets a say, and what commitments are being made to residents. It gives me somewhere to direct my concern beyond the abstract internet.

I give money away now, not later.

To churches and organizations, this year. I’m trying to make giving part of what I do with what I have, instead of something I’ll start once I’ve hit a number I keep moving.

I try to notice when the life vest is on.

Tech has trained me to keep asking: what’s next? What more? There’s always another milestone to reach, and reaching it mostly means setting the next one. I’m so used to trying to do better that what I already have can feel like it’s never enough.

Some of that uncertainty is real. I’m writing about a technology that is changing my own job, too.

But if helping other people depends on my feeling completely safe, I can postpone it forever. The feeling of safe enough to help may never arrive on its own. I have to decide what enough looks like, and what I can give from where I am.

Otherwise, I could spend my whole career adjusting the vest.

I still feel torn when I close my laptop and open my phone. I want the industry I work in to take more responsibility, and I want the people I love to have more opportunity.

But “the system won’t fix it” and “I can’t do anything” are two different sentences. I catch myself using the first to excuse the second.

I’m in some of the rooms where decisions get made. I get time with these tools as part of my work. I have people I can teach, questions I can ask, and money I can share.

That’s enough to give me somewhere to start on Tuesday.