Mean Shift
The New Baseline Is Already Here. The Question Is Whether You've Noticed.
If you spend any time in the AI conversation, you’ve met the two loudest people in the room. The first is convinced we’re eighteen months from digital utopia: disease cured, poverty solved, consciousness uploaded to the cloud. The second is equally convinced we’re eighteen months from extinction: job markets collapsed, truth dissolved, Skynet online. Both speak with absolute confidence. Neither knows.
Here’s what I find fascinating: the utopians and the doomers share the same underlying assumption. They both believe AI is a single, dramatic event with a clear before and after. A line in the sand. A moment when everything changes.
I don’t think that’s what’s happening. What I think is happening is quieter, more pervasive, and in many ways more consequential. We’re not witnessing an event. We’re witnessing a mean shift: a move to a new baseline, a new average of what’s possible and how work gets done. The ground beneath us is rising, and most people haven’t looked down yet.
The Data Says “Shift,” Not “Revolution”
The numbers tell a story that fits neither the hype nor the panic. Gallup reports that 45% of U.S. workers now use AI at work at least occasionally, with 23% using it frequently (Gallup, 2025). Microsoft Research finds that 75% of knowledge workers are using AI in some capacity, with top users saving a full workday per month (Microsoft Research, 2025). PwC’s global survey shows that 75% of AI users report productivity and quality gains (PwC, 2025). And ChatGPT sits atop the App Store’s productivity category in most countries worldwide (Similarweb, 2026).
But the aggregate productivity gain? The St. Louis Fed measured it at 1.1% economy-wide (St. Louis Fed, 2025). McKinsey sizes the long-term opportunity at $4.4 trillion, but concedes that only 1% of companies are “mature” in AI deployment (McKinsey, 2025). The Peterson Institute explicitly frames the research as “still in the first inning” (PIIE, 2026).
That’s not a revolution. That’s a mean shift in progress. The baseline is moving, but it hasn’t landed yet. Nobody knows exactly where it settles. And anyone who tells you they do is selling something.
Proof of Concept: One Pastor’s Workflow
I’m not a Silicon Valley developer. I’m a pastor with a congregation of about 800 people, a PhD student studying AI ethics at Southern Seminary, a husband, a father of four, a nonprofit board member, and the founder of an AI company. My constraint has never been ideas. It’s time. There is never enough of it.
AI has already radically changed my workflow, and I want to be specific about how, because the specifics matter more than the abstractions.
My Tesla’s Full Self-Driving capability has turned my commute from dead time into productive time. I use a Plaud Note Pro to dictate thoughts, article ideas, and sermon reflections while driving. That audio gets transcribed automatically. The transcript gets fed into an AI model that I’ve trained on my voice, my theological framework, and my writing style. That model outlines and drafts articles based on my dictated ideas, which I then edit, refine, and publish.
What used to require three to four hours of sitting at a desk now happens in recaptured margin. The ideas were always there. The time wasn’t. AI didn’t give me better ideas; it gave me a way to capture and develop the ones I already had, in time I was already spending on something else.
That’s one person. One writing workflow. Now multiply that across every knowledge worker, every industry, every bottleneck where the constraint is time and the redundancy is in execution, not in thinking. The tools are already here. They’re accessible. They’re not theoretical. Micro-level studies show 14 to 50% task-time reductions, with the largest gains going to lower-skilled workers (Law & Economics Center, 2026; Alex Imas, 2026). AI-exposed sectors are already generating three times the revenue per employee (PwC AI Jobs Barometer, 2025).
The mean is shifting. The question isn’t whether the technology exists. It’s whether you’ll use it.
Practical Wisdom: Find the Bottleneck, Build the Solution
Here’s my advice, and it’s simple enough to fit on an index card: know your industry’s bottlenecks. Find the redundant tasks. Build solutions with AI tools. Your solution might solve other people’s problems too.
Ankur Tyagi, a dev tool marketing expert, put it well: “Pick one workflow that hurts today. Use AI there until it clearly saves time” (Ankur Tyagi). That’s the right approach. Not “adopt 120 AI tools” (the internet is drowning in those lists). Not “replace your entire workflow overnight.” Pick the one thing that eats your time, apply AI to it, measure the result, and iterate.
A backend engineer named Rajat Sharma described his own arc honestly: he used AI aggressively early on, gained productivity, but noticed “weaker intuition, shallower reasoning.” His recalibration? “AI became my junior. Ownership stayed with me” (Rajat Sharma). That’s wisdom. That’s the posture.
McKinsey’s own research confirms it: high-performing organizations don’t just add AI tools to existing processes; they redesign workflows around AI (McKinsey, 2025). Addition is easy. Redesign is where the value lives.
The Warning: Don’t Confuse Output with Wisdom
Here’s where I put on my pastor hat, because someone needs to say this clearly.
Harvard Business Review reported in February that AI doesn’t reduce work; it intensifies it (HBR, 2026). Early adopters report higher work intensity despite efficiency gains. The ceiling of what’s expected rises with the floor of what’s possible. And HBR’s March piece on “AI brain fry” found that productivity actually drops when workers juggle four or more AI tools (HBR, 2026).
More output is not the same as better output. Faster is not the same as clearer. Productivity is not the same as fruitfulness.
This distinction matters theologically, not just practically. Scripture draws a sharp line between activity and fruitfulness. Jesus didn’t say “by their productivity you will know them.” He said “by their fruit” (Matthew 7:20). Paul’s list in Galatians 5 of what the Spirit produces in a human life: love, joy, peace, patience, kindness, goodness, faithfulness, gentleness, self-control. Not one of those is an efficiency metric. Every one of them requires the kind of slow, relational, embodied presence that no AI tool can accelerate.
The mean shift is real. The tools are powerful. I use them every day, and I’m building more. But I refuse to let the logic of optimization colonize every corner of human life. Some things should be slow. Some things should be inefficient. A conversation with your child, a prayer in the dark, the slow work of sanctification: these resist acceleration because they were never meant to be optimized. They were meant to be lived.
Stewardship, Not Anxiety
Genesis 1:28 gives humanity a mandate: be fruitful, multiply, fill the earth, subdue it, exercise dominion. Every tool since the first sharpened stone exists to help image-bearers fulfill that mandate. AI is the latest, and arguably the most powerful. But the mandate hasn’t changed. The purpose of the tool is to serve the purpose of the one who wields it.
The posture we need isn’t fear-based. It isn’t hype-based. It’s stewardship-based. Realistic about what these tools can do. Honest about what they can’t. Clear-eyed about the fact that nobody, including me, knows exactly where the new baseline lands.
What I do know is this: the mean is shifting. The people who understand their work deeply enough to identify the bottlenecks, apply the tools wisely, and retain ownership of the thinking will thrive. The people who either ignore the shift or outsource their judgment to the machine will not.
That’s not a prediction about technology. It’s an observation about human nature. And human nature, being what it is, hasn’t changed since Genesis 3.
The tools are new. The calling isn’t. Use them well.