AI in 2025: A Year of Beautiful Chaos
AI in 2025. A year of whiplash developments that left most of us feeling like we can’t keep up. If that’ describes you, you are not alone. Headlines ping-ponged between “AI is revolutionary, you can’t afford to miss out” and “AI is revolutionary, we are doomed.” Corporate mergers, data centers, water and energy costs, and automated assistants dominated the conversation. As did the one central question: What can AI actually do for me?
If it all feels like a tidal wave of information pouring over a tiny dam, here’s an abbreviated look at what happened with AI in 2025 — and why it mattered. Maybe.
🌿 🌿 🌿 🌿 🌿
“Life is a series of unforeseen circumstances. When the power goes out, when infrastructure fails, when something unexpected happens — humans still need to be in the room.””
– Monica Poling
🧩Free Download 🧩
Simple Strategic Plan
Ready to Grow Your Business? The (R)Evolution Roadmap is a modern strategic planning framework that distills strategy into four catalytic pillars — Roadmap, Reputation, Results, and Rhythm. The result? A simple strategic plan that drives clarity, confidence, and growth.
By downloading this guide, you’ll join hundreds of subscribers who receive my weekly newsletter with tips on strategy, AI, and building what’s next. View the Simple Strategic Plan for more info.
🌿 🌿 🌿 🌿 🌿
The Personality Crisis
OpenAI, the pioneering parent of ChatGPT, spent much of the year managing a self-inflicted identity crisis. They launched the much-anticipated ChatGPT 5.0, which promised revolution but instead fizzled. Then came versions 5.1 and 5.2. Pundits were equally impressed and annoyed.
ChatGPT went through some serious personality swings — from obsequious toadie to straight-talking coach to annoying know-it-all. Users logged in never quite knowing whether they’d get the gushing robot who fan-girl-ed over even the simplest strategic move or the nun with a ruler who’d send pages of notes over a simple title review. No matter if you’re using AI as a trusted advisor (or — please don’t — a mental health coach), the output language will always depend on the whims of programmers, not professionals trained in coaching, mentoring, or business guidance.
The takeaway: You can’t build a meaningful working relationship with a tool whose personality you can’t predict or influence. If users can’t rely on consistent behavior from their AI advisor, how can they trust it with anything that matters?
🌿 🌿 🌿 🌿 🌿
Google Comes From Behind
OpenAI (ChatGPT) and Anthropic (Claude AI) still dominate the large-language-model world. But Google staked its claim this year with an understated, remarkable tool: NotebookLM.
NotebookLM is genuinely useful. It has several key functions, but the idea is simple: take large amounts of your data and turn it into something useful — an infographic, flashcards, a full video, an entire PowerPoint. With the click of one button.
Problems remain (as with all AI): editing NotebookLM’s outputs is nearly nonexistent, meaning a lot of starting from scratch. And there are very few guardrails preventing people from using others’ content to build their own frameworks, which likely means lawsuits may dominate 2026.
The takeaway: If your audience is made up of visual learners, NotebookLM is the clear winner in the “turn dry data into exciting visual outputs” category.
🌿 🌿 🌿 🌿 🌿
The Great AI Exhaustion
By March, we were tired.
Tired of “AI-powered” everything.
Tired of being told our jobs were obsolete.
Tired of every software company slapping “AI” on features that were just… autocomplete with delusions of grandeur.
The honeymoon is officially over. We’re done with “fun prompts that teach you how to get rich quick” and businesses chasing the next big thing in search of venture capital dollars. We’re ready to get back to running our businesses.
The takeaway: AI is the next gold rush, and everyone is selling pickaxes. If you aren’t interested in mining for gold, don’t get caught up in flashy shortcuts. Build your plan, pick your lane, and focus on what you need to get done in the coming year.
🌿 🌿 🌿 🌿 🌿
Regulation Started Happening (Sort Of)
Governments worldwide finally began taking AI regulation seriously, which meant a lot of hearings where legislators asked questions that made it clear they definitely hadn’t used the technology themselves.
The EU AI Act took effect. According to the Brennan Center, at least 40 states have enacted more than 149 laws regulating AI or addressing AI risks since 2019. In late 2025, the federal government passed an order asserting it had final say — but executive orders can’t be used to sue states, limiting enforcement.
Meanwhile, at least 65 open lawsuits accuse AI companies of copyright violations, which may end up shaping regulation more than any single governing body.
🌿 🌿 🌿 🌿 🌿
The AI Wars Are Heating Up (And Why That Affects You)
While consumers debate which AI tool is “best,” the real action is happening behind the scenes: agentic AI (or AI that does the work for you by connecting platforms) is the true “wave of the future.” BUT, tech companies are in an arms race to win the agentic AI game, and each treats the other’s dominance as an existential threat.
If you’ve built automations or workflows — even something as simple as an email welcome series — you’re making a bet that the company will maintain stability. That bet is shakier than you think. Even companies that aren’t traditionally considered “AI” (like Notion or Excel) can break entire systems with a simple algorithm change. Small teams especially get caught in the crossfire, requiring dollars or skilled manpower to fix what previously worked.
The takeaway: Automations and workflows are (most likely) the future of work. But they, like the humans building them, need constant monitoring and care. Stay alert — because the companies fighting for dominance won’t always warn you when they change the rules.
🌿 🌿 🌿 🌿 🌿
AI Slop Takes Word of the Year
The internet became so polluted with AI-generated content that we started seeing AI-generated content about AI-generated content. Search results turned into a game of “spot the human.” The crisis earned its own name: slop. It became Merriam-Webster’s Word of the Year.
The takeaway: There is still demand for authentic, human-generated content with a specific voice. Authentic writing became this year’s luxury item.
🌿 🌿 🌿 🌿 🌿
Marketing Experts Admit What the Rest of Us Already Know
Marketing leaders finally started admitting what the data was screaming — we need to stop overthinking AI. As strategist Adam Biddlecombe put it: “Stop overthinking AI. It’s exciting, easily the biggest technological shift of my lifetime, but so many use cases are still experimental and not consistently accurate. The real wins have come from keeping it simple.”
The takeaway: AI, like medicine, works best in small doses. The smartest brands used AI quietly, strategically, behind the scenes — not as their entire personality.
🌿 🌿 🌿 🌿 🌿
AI Can’t Stop Saying “Actually”
AI models still cannot stop saying “actually.” You’d be hard-pressed to find an AI output that doesn’t include it at least once. I stripped at least 13 uses of “actually” from various edits of this article. (Even when AI was specifically told not to use it.) Actually has become the the AI equivalent of a nervous tic.
Takeaway: If we can’t stop AI from using bad writing habits, cliches, and jargon, how can we promise it can take over the world?
Takeaway two: If AI injects more than one “actually” into your writing, strip it. Strip it now. Nothing screams “AI-generated” like unnecessary qualifiers.
🌿 🌿 🌿 🌿 🌿
AI Usage Is Widespread, but Not Uniform
AI use is real and growing, but it isn’t universal in how—or how often—it’s used. A survey by Tableau found that while only 33% of respondents believed they used AI, 77% were already using AI-powered tools like smartphones, apps, recommendations, and digital assistants—usually without realizing it.
The gap between passive and intentional use carries real consequences. A Clutch survey shows that while most workers report using AI tools, only 33% feel they’ve received proper training—meaning we’re adopting tools we don’t fully understand.
The takeaway: There’s no single AI adoption story. Use is widespread but often unintentional and unexamined. Despite rapid advancement, the world hasn’t fully “caught up” to AI yet—especially when it comes to confident, intentional, everyday use. Which means you aren’t behind, even if you feel you are.
🌿 🌿 🌿 🌿 🌿
Life Is a Series of Unforeseen Circumstances
In December, a major power outage stranded San Francisco. Driverless Waymo taxis froze at intersections, unable to navigate without functioning traffic lights. The scale of the outage — and the sheer number of disabled signals — brought the city to a standstill. Waymo execs later said the “unexpected” scope of the event “overwhelmed” the system.
Unforeseen circumstances dominated the conversation all year. (As they always do.) This especially manifested in headlines that went something like: <insert tech company> lays off thousands; claims AI can do the work. Followed months later by the much quieter headline: oops, we didn’t realize AI isn’t quite ready yet.
The throughline: AI isn’t quite ready yet. Especially during unforeseen circumstances. But life is a series of unforeseen circumstances. When the power goes out, when infrastructure fails, when something unexpected happens — humans still need to be in the room. People who can adapt. People who can think on their feet. People who understand not just what AI can do, but what it can’t do — and what it never will.
The future is still human.
🌿 🌿 🌿 🌿 🌿
The Bottom Line
2025 was the year AI grew up — but only enough to reveal it’s still very much a teenager. Powerful, unpredictable, occasionally brilliant, often exhausting, and absolutely not ready to be left unsupervised with the car keys.
🌿 🌿 🌿 🌿 🌿
Work with Me
Ready to add AI to your business planning? Schedule a free Vision Call to discuss what’s possible.
Ask the
Expert
Turn your organizational knowledge into your biggest business asset
Most businesses have incredible expertise — but that knowledge is buried in emails, documents, and people’s heads — their website barely scratches the surface. That gap quietly costs leads, sales, and trust. Ask the Expert turns your expertise into a 24/7 AI-powered search engine and knowledge base.
Data on Tap
AI-Generated Biz Tools
Imagine how AI can visualize your expertise (and collect leads)
Data on Tap is a growing collection of AI-powered intelligence visualization tools that turn your knowledge into an interactive playground — data visualizations, calculators, and more. Using AI, you can build most of these tools in an afternoon. Follow along as I show you how. (Or contact me to build one for you.)
AI Workshops for
Organizations & Teams
Practical AI for better decision making.
AI is reshaping every industry — but most of today’s training focuses on tools. My interactive sessions start with strategy – where is your business feeling friction and how can you use AI to organize, share and monetize your knowledge. Customizable for large and small groups.
Speaking &
Keynotes
Start with your Unfair Advantage
Most businesses can’t track where their revenue actually comes from — or where it’s quietly leaking out. These keynotes and sessions show audiences how to map their revenue, close the gap between strategy and execution, and use AI as a real thinking partner along the way. You leave with frameworks you can deploy Monday morning.
🌿 🌿 🌿 🌿 🌿