By John Barrows | September 2026

I have had about a dozen conversations recently with founders, CROs, senior executives, and individual reps that all ended up in the same place. Different industries, different levels, different companies. Every single one of them landed on the same question: what the hell are we actually doing? Not “how do I use AI better.” Why. And to what end. I did not have a great answer for them, and I am not sure I have one for myself right now either.

Why the AI anxiety is real and shared

The pattern I keep seeing is this. Someone wants to talk about business. Ten minutes in it stops being about business. They start describing what they are actually dealing with: the pace, the pressure to have an AI answer for everything, the feeling that no matter how much they read or how many tools they try, they are behind. Then I tell them what I am dealing with, and it is the same list.

This is not imposter syndrome. It is not a skills gap. It is a genuine and rational response to a situation that is moving faster than anyone can responsibly process. The people I am talking to are not naive. They are experienced, successful, and paying close attention. And they are exhausted in a way that feels different from the normal kind of burnout, because normal burnout comes from doing too much of something you understand. This is something else.

The 10% problem

Here is what pushed me over the edge on this. Anthropic’s head of alignment science, one of the people whose actual job is to keep this technology safe, said publicly that he thinks there is a greater than 10% chance AI kills all of us within the next decade. In the same period, a researcher who had worked at both Anthropic and OpenAI resigned and said publicly that neither company is acting responsibly.

Read that again. That is not a critic or a politician talking. That is the person inside the building whose job is to prevent the worst outcomes, saying openly that the odds of the worst outcome are not negligible. If a plane had a 10% chance of going down, nobody would board it. We are treating this like a productivity tool. The people building it are talking about it like it is nuclear.

The efficiency phase and what comes after it

Here is what I actually think is happening, and it is the part that concerns me most. We are still in the efficiency phase of AI. It makes what we already do faster and cheaper. That is the phase most of us are in right now, and it is the phase most of the training and advice is designed for.

The results phase is next. That is where AI does not help you do the work. It does the work for you. When that arrives, a lot of the roles people are frantically trying to protect will not need a human in them. Not because the humans were not good at the job. Because the job itself no longer requires one. That is a different conversation than “how do I prompt better,” and I do not think we are having it seriously enough yet.

The cost nobody is talking about honestly

Every time I open an AI tool, there is a voice in the back of my head that my wife put there. She is an environmental scientist who works on climate and sustainability with communities that are living with the real consequences of all this. The impact of data centers is not abstract in our house. The International Energy Agency reported that data center electricity use jumped 17% in one year, with AI-focused centers growing even faster, and the whole thing is on track to double by 2030 (IEA, April 2026).

The deal I made with myself was simple: if I am going to use it, the good has to outweigh the harm by a lot. That deal is getting harder to keep. And I want to hold the other side of it too. Moderna and Merck recently announced that their personalized melanoma vaccine passed a Phase 3 trial, a custom shot built for each patient’s tumor that cut the risk of cancer coming back. That does not happen without the same computing power that is chewing through the grid (CNN). I am holding both of those things at the same time and I am not going to pretend it is comfortable.

Where the hope actually is

I do not want to leave it at the hard part, so here is where I actually see light. The same tools that are flattening traditional jobs are also collapsing the cost of starting something. U.S. business applications hit 5.6 million in 2025, up 24% since ChatGPT launched, and the pace is still climbing (Census Bureau, July 2026). Startups are launching with smaller teams because one person with the right skills can now do what used to take five or ten.

I personally believe the future is entrepreneurship. Not everyone becoming a founder of a billion dollar company, but people with specific skills who know how to run agents and plug themselves into different businesses and different problems. That is the job of the future as far as I can tell. It is why I am teaching my daughter how to build right now, at 15. I want her to know how to generate her own income and be financially independent before she ever needs to be, so her future does not depend on a college degree holding its value or a company keeping her until the day AI can do her job cheaper.

What you can actually control: E-A-T

When my daughter was about six or seven, we came up with something together that I have talked about a lot since. E-A-T. Effort, Attitude, and how we Treat people. Those are the only three things you fully control. Not the market, not the model releases, not what some lab decides to ship next quarter.

You cannot out-learn a technology that is improving weekly. What you can do is show up with real effort, keep your attitude in a place where you are still curious instead of scared, and treat the people in your life well while everything else shakes out. We still do a secret handshake every morning before she leaves for school. It is a reminder for both of us that the noise is the noise, and the work is the work.

The one lever most people never pull

Focusing on what you control includes the one lever most of us never use. If the people building this technology are telling us it might be an existential risk, then it needs to be regulated like one, and it needs to happen fast. Right now Congress is treating AI like a tech policy issue that can wait. It cannot.

You do not need to be an expert to say that. You need to be a constituent. Congressional staff track what people call and write about, and the issues that light up the phones are the ones that get in front of the member. Here is how to do it in about five minutes.

Go to actonai.org, pick the issue that matters to you, enter your location, and it connects you to your representatives with a message you can edit. If you would rather go direct, find your House member at house.gov and your senators at senate.gov. Keep the message short and personal: who you are, where you live, what you do, and that you want them to prioritize AI safety regulation now, before the results phase gets here. If you want to see what is actually moving in Congress first, Legisletter’s AI regulation tracker lists every bill and vote.

If a few thousand people do that, it registers. That is not complaining. That is doing something.

Frequently asked questions

What is AI anxiety and why are so many sales professionals experiencing it?

AI anxiety is the specific kind of stress that comes from operating in a field that is changing faster than anyone can responsibly track, while feeling pressure to have an answer for all of it right now. Sales professionals are experiencing it because the technology is directly threatening to automate significant parts of the job, because the pace of change makes it impossible to feel genuinely caught up, and because the signals coming from the people building the technology suggest the stakes are higher than most public conversation acknowledges. It is a rational response to a genuinely uncertain situation, not a personal failure.

What is the difference between the efficiency phase and the results phase of AI?

The efficiency phase is where most organizations are right now. AI makes existing work faster and cheaper. It helps a rep research faster, write faster, and follow up more consistently. The results phase is what comes next: AI does not assist the work, it does the work. In the results phase, many roles that currently require a human to execute will be automated entirely, not because the humans are underperforming but because the underlying task no longer requires human judgment to complete. The distinction matters because the advice and training that helps in the efficiency phase does not prepare people for what the results phase demands.

What does the E-A-T framework mean?

E-A-T stands for Effort, Attitude, and how you Treat people. It is a framework for staying grounded when external circumstances are outside your control. Effort means showing up consistently and doing the work regardless of what the market or the technology is doing. Attitude means staying curious rather than scared, which is genuinely harder than it sounds in an environment that keeps changing. Treat people means maintaining the quality of your relationships and how you show up for the people around you even when everything else is uncertain. These are the three things nobody and nothing can take from you.

Why does AI regulation matter for people in sales?

The people building the most powerful AI systems have said publicly that the technology poses serious risks if it is not developed responsibly, including at least one senior researcher citing odds of catastrophic outcomes that most people would consider unacceptable in any other context. If those risks are real, the pace at which this technology gets deployed into the economy without safety requirements in place is a problem that affects everyone, including sales professionals whose roles are already being reshaped by it. Regulation is how societies manage technologies with large-scale consequences, and AI is overdue for the same treatment.

What is the Future Founder Project?

The Future Founder Project is an initiative built on the belief that entrepreneurship is the most durable path forward for young people entering an economy where AI is eliminating traditional entry-level roles faster than new ones are being created. The premise is that a young person who knows how to generate their own income, run agents, and build something is more resilient than one whose career depends on a credential or an employer deciding not to automate their position. The project focuses on getting tools and resources into the hands of young people who have the drive to build something before they ever need to.

Keep going

The sales fundamentals that matter in every phase of this, efficiency and results alike, are at the core of Filling the Funnel and Driving to Close. See what teams are building at learn.jbarrows.com/pages/results.

John Barrows helps sales leaders decide whether to replace or rebuild their teams for the AI era. For 25+ years he has worked with the world’s most demanding sales organizations, including Salesforce, LinkedIn, Google, Amazon, and Okta, building the frameworks that became Filling the Funnel and Driving to Close. He is the host of Make It Happen Mondays, author of I Want to Be in Sales When I Grow Up, and an LP at GTMfund. His training programs are available at learn.jbarrows.com.