Eric Levitz wrote a Vox piece called “AI’s threat to white-collar jobs just got more real” that (metaphorically) equated the rise of AI to the month before COVID, a time that would change us, physically and psychologically, forever. Except this time, it’s not a virus. It’s technology that’s learning to do our jobs, and many of us are still casually going about our business like nothing has changed.
This argument is bolstered by Anthropic CEO Dario Amodei, who published a 20,000-word essay titled “The Adolescence of Technology” about the risks AI poses to the job market. Here’s a quick overview with some additional insight.
In the 1970s, ATMs initially reduced the number of tellers, but more branches opened due to lower branch-building costs, which led to more tellers and the opportunity to expand their skill sets into relationship banking, sales, and customer service. Going back further, when mechanized farming shifted us from an agrarian to an industrial economy, the number of farmers plummeted, but they transitioned to blue-collar factory workers, mechanics, salesmen, and operators.
But unlike early computing or mechanized farming, which have lasted decades, maybe even over a century, depending on whether you start with the horse-powered or combustion eras, AI is progressing at a rate many humans find impossible to match. That speed of progress is threatening ALL cognitive work at once, making it hard for people to simply switch to an adjacent field as tellers or farmers did. Amodei has even gone on record stating that the once-coveted coder job is rapidly being replaced by AI; 90% of the code at Anthropic is written by Claude Code.
What’s more, AI companies are making models smarter, faster, cheaper, and more accurate every day. Remember when AI-generated hands had six fingers instead of five? That bug was fixed, as have millions of other bugs over time. I said in 2024 that this is the worst AI will ever be and was dismissed by many people whose view of AI revolved around a sense of control and its “limited” application. That worldview has clearly shifted in less than two years.
Amodei also discusses the wealth gap and issues a stark warning: right now, most people tie their sense of purpose to their jobs. If AI takes over most work, society needs to find new ways for people to feel valued and fulfilled. The underlying threat is the potential creation of a permanent underclass whose skills have been automated out of existence. And the irony of it all is that he’s betting on AI to help us figure out what that looks like.
Knowledge workers still have an edge
The Vox piece, and most of the AI-and-jobs conversation, spends a lot of time on what’s being disrupted and very little time on what makes humans irreplaceable. Let’s use me as an example. I’m a knowledge worker with a deep sense of purpose. For 25 years, I’ve been a problem-solver and solution-finder, and yes, so is AI. I’m also a strategist and advisor who can build and develop teams and the systems they work in. And truthfully, AI can do that too, but not all of it.
The one major thing AI can’t do is feel.
AI can process sentiment or detect tone. It can predict what you’re likely to say next based on patterns in billions of conversations, but it does not have the capacity to sit across from someone and discern that the words coming out of their mouth aren’t the real problem. That’s emotional intelligence. And right now, it’s the single most important competitive advantage any professional can build.
We’ve been culturally wrestling with this for decades. In the sci-fi classic I, Robot, Will Smith’s character, Detective Spooner, asks the question: “Can a robot write a symphony? Can a robot turn a canvas into a beautiful masterpiece?” and Sonny the robot inquisitively replies, “Can you?” Sonny’s innocent retort sparks a distinctive debate. It’s not whether anyone or everyone can create art (just contemplate the artistry of Xania Monet, China Styles, or the scribbling in a two-year-old’s coloring book). Instead, it’s whether a machine can understand why a piece of music makes you cry, why a leader’s silence in a meeting says more than their words, or why a customer’s hesitation on a call means they’re not objecting to the price but may not yet trust you.
Emotional intelligence is tied directly to the human experience. You develop it by living, failing, leading, being led, navigating loss, building relationships, and learning to read what lies beneath what people say. It has no capacity to experience love in any form. With no body to feel stress, pain, loss, joy, or fear, how can it authentically capture the true stakes of human existence? Neo from The Matrix movie doesn’t defeat the system by being faster or more powerful than the machines. He wins because he makes a fundamentally human, irrational, love-driven choice that the system couldn’t compute. The machines understood everything about human behavior except the part that actually drives it.
The AI Gap Is Your Career Strategy (At least it’s mine)
AI has the data and speed. Humans have the edge in emotional intelligence and sound judgment, though we don’t always use these skills perfectly.
Before we go further, let’s retire “soft skills” from our vocabulary. There’s nothing soft about them. The ability to navigate conflict, coach a team through uncertainty, negotiate without formal authority, read a room, give constructive feedback, and hold space for a difficult conversation are the hardest skills in business. Let’s reframe and call them what they are: power skills.
Here are the five power skills to master:
- Emotional discernment. This is the ability to understand what drives people beyond algorithmic logic. The CEO who says the sales team needs a reorg might actually need to hear that her real problem is that the sales team doesn’t feel like they have enough autonomy within the current system. AI will hand you the data. Emotional discernment tells you what the blocker truly is.
- Navigating tension. I built a framework and workshop called the Four Tensions of Change, which helps leaders identify the type of change challenge they’re experiencing and apply the right tools to receive feedback, coach for autonomy, negotiate among competing stakeholders, and resolve conflict in ways that strengthen relationships. AI can hand you frameworks like mine to study, but reading the room while you’re standing in it is a different skill.
- Contextual judgment. Knowing when to push and when to pause, when to follow the data and when to override it because something doesn’t feel right, or when to speak and when to listen are learned skills that distinguish effective leaders from everyone (and everything else). These skills are developed through exposure to complexity over time. AI can tell you strategies to use, but the external cues are not something it can teach you or discern for itself.
- Trust-building under pressure. You absolutely cannot automate trust. You earn it by showing up consistently, telling hard truths kindly, delivering on promises, and demonstrating that you understand what’s at stake for the other person and for you. That’s what makes a true leader. AI trust collapses the moment one answer turns out to be wrong or when the stakes of potentially getting it wrong are too high, which is exactly why humans must stay in and over the loop.
- Cultural fluency. Understanding how an organization actually works beyond the org chart is imperative. Power dynamics, unspoken norms, and even the history that shapes every decision are invisible to AI. Can AI tell you that your CMO may hesitate to share bad news or experiment with new ideas because she was burned by the last CEO who valued quick wins over long-term performance? Nope! That intel lives in people, not in databases.
The Bottom Line
Power skills aren’t supplementary skills. In the AI era, they are the job. The Vox piece ends with a powerful line: “If celebrations of the singularity are premature, preparations for something like it are now overdue.” I think that’s exactly right. But I’d add this: preparation isn’t learning to code or mastering the latest AI tool, but rather it’s urgently becoming the kind of leader, thinker, and operator that no machine can replicate.
Data has been the king of the playing field for a decade. AI just crowned itself emperor. But the throne that matters most belongs to whoever can read what the data actually means for people and to lead them through what comes next. We’re watching technology do meaningful parts of our work, but not perfectly, at least not yet. But the gap between ‘not quite’ and ‘good enough’ is closing faster than most people realize, and ‘good enough’ has a way of winning.
Maybe one day our inventions will create a utopian society that frees us from the economic pressures of today. Until then, if there was ever a time to get better on purpose, it’s right now.
Karen McFarlane is a strategic B2B marketing executive, board director, and venture builder with more than 25 years of experience, including 14 years as a fractional CMO and CMO-in-Residence helping companies scale smarter. As founder of Kaye Media, she partners with high-growth startups and mid-market B2B companies that need senior marketing leadership to accelerate growth, refine positioning, and develop next-generation leaders. Known for bringing the strategic mindset of a CMO and the operational discipline of a COO, Karen helps organizations strengthen their message, align their teams, and build systems that turn ambition into measurable growth.
Whether guiding startups like Fivetran and CareAcademy, repositioning SaaS challengers like Smartling and Creative Virtual, or supporting global leaders such as Amazon, Twilio/SendGrid, Braze, and Knowable (a LexisNexis company), Karen has delivered outcomes that change trajectories. She has scaled pipelines to seven figures, cut go-to-market cycles from 12 weeks to 6, and reduced customer acquisition costs by 20% while accelerating pipeline velocity. She has repositioned brands to category leadership, multiplying marketing-qualified leads fivefold and translating into millions in new revenue opportunities. She has also led AI-driven product launches with Creative Virtual and Knowable, combining category strategy, messaging, and demand generation to bring AI-powered innovations to market.
Karen’s leadership extends beyond client work. She served as president of the American Marketing Association New York and AMA’s Professional Chapters Council, leading volunteer boards responsible for strategy, operations, and leadership development nationwide to strengthen infrastructure and increase member value. She also develops ventures at the intersection of innovation, culture, and business—from launching Really Good Intros, a visibility platform for overlooked talent, to co-hosting The E Word podcast, a show exploring leadership, culture, and systems of power.
A published author, speaker, and recognized thought leader, Karen has written influential industry reports on inclusive marketing and organizational strategy and was named one of Crain’s Most Notable in Marketing & PR. She serves on the board of the Hudson Valley Credit Union ($8B in assets) as Chair of the Compensation and Benefits Committee, and on the board of AMA New York as Secretary/Treasurer.
Karen is an NACD© Certified Board Director and holds an MBA in Marketing and a BS in Communications from New York University, and has completed executive education in governance, negotiation, executive compensation, and entrepreneurship through programs at Columbia Business School, Wharton, MIT, and Cornell.
Subscribe to her LinkedIn newsletter, Better on Purpose.