The Word Up! Blog

Humans must stay in the picture. Not because AI is bad. Because the picture stops being a picture without us. The room, the eyes, the breath, the recovery, the story you can only tell because you lived it. None of that is a feature you can ship. All of it is the skill that just became the most valuable thing you have, and it is only built one way: live, with other humans, on purpose.

I am a creative-technical founder. MIT-trained. I build with AI every day at Business 4 Good, where we put the human intelligence into artificial intelligence. I ship with it, debug with it, prototype with it, and I will keep doing all of that. I also drill at Speaker Skills Academy, where I have been working on personal narrative pieces using a framework called S.A.M., Snapshot, Affect, Mirror. From those two seats, the math is becoming obvious. AI is getting better at almost everything that used to require a human. There is one category where it is not catching up. Live, in-person speaking is the skill that is appreciating in value, and most of us are still practicing it the wrong way, or not at all.

This is the case for taking the live work seriously, before everyone else does, and the case for keeping real people at the center of how you build it.

Why Communication Just Became Your Most AI-Proof Skill

The major reports on the future of work all say the same thing in slightly different language. The World Economic Forum’s Future of Jobs Report 2025, drawing on data from over 1,000 employers across 55 economies, found that 39% of workers’ core skills will change by 2030, and named “leadership and social influence” among the top 10 fastest-rising skills. That is communication, in plain language. LinkedIn’s 2025 Skills on the Rise list puts AI Literacy at number one and verbal communication on the same list, with the rationale that effective verbal communication is essential for inspiring confidence in your ideas and expanding your influence.

The reason is simple and it is not nostalgic. AI is good at outputs. The skill that is appreciating is not output. It is reading, adjusting, and landing in real time, with real humans, under real pressure. That is a different muscle.

What AI Can’t Do (And Probably Never Will)

I am pro-AI. I use it constantly. I am also clear-eyed about where it stops. Watch where it stops and you are looking at the skill that just became scarce.

AI drafts your slides beautifully. AI cannot read the room when slide three lands flat and your CFO checks her phone. AI rehearses your pitch in voice mode. AI cannot adjust when an investor asks you a question you did not prepare for, in a tone that tells you she already has doubts. AI writes a flawless toast. AI cannot stand at the head of the table on Thanksgiving and feel, in your body, that it is time to wrap. The body language work, the recovery work, the timing work is yours. The Silent Power of Body Language in Speaking made the case for one piece of this. Guided Chaos made the case for another. The pattern across all of them is the same: AI handles the prep, you handle the live.

This is not a hedge against AI. It is a clean read on the division of labor. AI gets faster every quarter at the prep work. The live work is yours, forever, if you build it.

The Ethics Problem Nobody’s Telling You About

Here is the part most AI-and-the-future-of-work articles skip. AI does not generate from nothing. It generates from the data it was trained on, which is the internet, which is mostly written by a narrow slice of humanity. Researchers have been documenting this for years. Harvard Business Review calls it baked-in bias. Brookings Institution has published guidance on detecting and mitigating it. The pattern is clear. Training data carries every blind spot, every default, every cultural assumption baked into who got to write the most words on the most platforms over the last twenty years. When you let AI rehearse your pitch, polish your toast, or coach your story, you are not getting neutral feedback. You are getting feedback shaped by whoever happened to dominate the training data. That is not malice. That is math.

It matters in speaking specifically because speaking is where you put your voice into the world. If the voice you are training is the AI’s averaged-out voice, you are not getting better at speaking. You are getting better at sounding like the median of the internet. That is a real cost, and it is not theoretical.

The fix: Practice in front of real, varied humans who will tell you the truth. A coach from Brooklyn who has won speaking championships hears your story differently than a coach from Portland who has put TEDx speakers on stage. A community of dozens of voices, from dozens of backgrounds, on dozens of cadences, gives you something the model literally cannot. Corrective feedback from outside the average. And you can connect with all of these people weekly through SSA’s programming.

The ethics question and the practice question turn out to be the same question. Who is in the room when you build your voice? If the answer is “a model trained on the internet,” you have a problem. If the answer is “people, on purpose, with names,” you have a path.

The Practice Gap: Why Most Speaking Programs Miss the Point

If communication is the skill that just got more valuable, you would expect everyone to be racing to practice it the right way. Most are not. Most are doing the same thing they were doing in 2015. Reading a book. Watching a course. Joining a club that meets twice a month and gives you four minutes at the lectern.

Those things are not bad. They are insufficient for what the AI moment is asking of you. A book teaches you to think about speaking. A course teaches you the theory. A traditional club gives you a few minutes of stage time on a slow cadence. None of them puts you in enough live reps, in enough varied formats, with enough pressure and feedback from enough different humans, to build the muscle the room actually rewards.

Six Things Everyone Who Uses Speaking to Communicate Needs in the AI Era

Stack the sessions. Stack the skills. Six things every speaker needs. One place that actually delivers all six.

The Real Humans Behind SSA (And Why That’s the Whole Point)

Here is what every speaking app and AI coaching tool cannot put on their landing page. Real humans, with names, who happen to be excellent at this and also happen to be really cool. SSA was founded by Cathey Armillas and Marc Williams. Cathey is a TEDx speaker and one of the most sought-after TED Talk coaches in the country, currently coaching a TEDxPortland speaker to the stage this year. Marc is a Brooklyn native and a global speaking champion, with a wall of national and international titles to back it up. Neither one of them is a chatbot. Both of them will look you in the eye on a Wednesday afternoon and tell you the one thing about your delivery that nobody else has told you.

The coaching bench around them is the same story. Gina Riley, George Rivera, Karin Kusumakar, Alexina Atin Shaber, Teresa Younkin, Lefford Fate, and the broader community of coaches, every one of them an actual human with an actual point of view, an actual track record, and the willingness to bring it into the room. That is the real differentiator. AI can simulate a coach. SSA gives you the one who has lived through the moment you are about to live through, and who actually likes you.

This is the answer to the ethics problem and the practice problem at the same time. The voices in the room are real, varied, accountable, and warm. The feedback comes from people, not models. The community is built on the radical idea that humans get better at speaking by speaking with other humans, and that the humans should probably enjoy each other while doing it. That last part is not soft. It is the whole engine. Joy is what makes you show up next week.

What Drilling at SSA Actually Built for Me

I will give you one moment. Six months ago I sat in a Skills Drill working on a personal narrative piece using S.A.M., the Snapshot, Affect, Mirror framework SSA uses to help you find the load-bearing beat in your own story. I had told the same story a hundred times. In the drill I had to tell it in ninety seconds. Then sixty. Then thirty. Then with a constraint I would not have chosen for myself, with a coach watching, with a room of strangers listening.

Somewhere around the third repetition the story finally cracked open. The Snapshot got specific. The Affect got honest. The Mirror, the line that turns my story into something the audience could feel was about them, finally landed. Six months later, I can pull that story out at a dinner, in a pitch, on a Zoom, and it lands the same way. Not because I memorized it. Because I drilled it, with real people, who told me when it was working and when it was not.

AI can write your story. AI cannot drill it into you. Only humans can do that, and only humans should.

That is the moat. That is the whole reason I keep showing up. And it is the reason humans must stay in the picture, not as a sentimental afterthought, but as the load-bearing structure of how anyone gets actually good at this. The room is irreplaceable because the people in it are. Take the people out and you do not have a faster version of the room. You have no room at all.

Practice still wins. Show up and drill. With real humans, on purpose.

Word Up!

Join us at the next free Open Session

Monday, May 4, 2026 · 3pm PST / 6pm EST · Live on Zoom

RSVP for May 4 →

Or learn more about Open Sessions first