There are headlines you read, shake your head at, and move on from. Then there are headlines like this one: AI could kill all humans within the next decade.
It sounds like the plot of a sci-fi movie. Some of the people raising these concerns are building the technology.
Former Anthropic researcher Jacob Coxon recently resigned and publicly warned about the race toward increasingly powerful AI systems. Anthropic alignment researcher Evan Hubinger has said he personally puts the probability of AI killing all humans within the next decade at greater than 10%.
When the potential consequences are enormous, responsible leaders don’t wait for certainty before managing risk.
AI Risk Is Already a Workplace Culture Issue
I don’t build AI models but I build healthy workplace cultures, and the AI conversation is exposing a very familiar organizational problem.
A company discovers something that delivers extraordinary results. Everyone gets excited about its potential. Pressure builds to move faster. Competitors are in the race. Leaders don’t want to fall behind. Then someone raises a concern and suddenly they’re the person standing in the way of progress.
I’ve watched this scenario play out with high-performing toxic leaders for years. A leader generates revenue, delivers results, or possesses expertise the organization believes it can’t afford to lose. At the same time, employees tell HR that this person’s behavior is creating fear, burnout, turnover, or psychological harm.
Leadership sees the results. Employees experience the risk. The organization keeps tolerating the problem because the damage hasn’t become expensive enough yet. Then it does. Leaders cannot take the same approach to AI.
Organizations need the culture, systems, boundaries, and accountability to manage AI responsibly before the consequences force them to act.
The Biggest AI Risk for Most Organizations Isn’t Extinction
Forget killer robots for a minute. Organizations are already facing AI risks inside the workplace.
- Employees paste confidential information into public AI tools.
- Managers use AI-generated information to influence hiring and performance decisions without understanding how the output was produced.
- Teams automate work without adequately addressing bias, privacy, security, or accuracy.
Employees also start trusting AI-generated answers because the technology sounds confident, even when those answers are wrong.
And in the race for efficiency, organizations risk stripping away the human judgment, relationships, conversations, and critical thinking that make workplaces function.
What Leaders Can Do Right Now
Leaders must take responsibility for how AI operates inside their organizations. Here are four places to start.
1. Create Clear Ownership for AI
Someone must own AI governance.
Leaving ownership of AI solely to IT is a mistake. AI affects people, privacy, performance, customers, decision-making, risk, and workplace culture.
HR, legal, cybersecurity, operations, and business leaders need clearly defined responsibilities for how AI is selected, implemented, monitored, and used.
When everyone owns AI, nobody owns AI. Make accountability explicit.
2. Define Your Non-Negotiables
Don’t wait for an AI-related incident to decide where the boundaries are. Define where human judgment is mandatory.
- Hiring
- Terminations
- Performance evaluations
- Disciplinary decisions
- Confidential information
- Employee data
Leadership needs to determine what AI can support, what requires human review, and what AI should never decide. Set those boundaries now, before speed and convenience set them for you.
3. Train People to Challenge AI, Not Just Use It
Teaching employees how to write prompts isn’t AI literacy. Employees need to understand hallucinations, bias, privacy concerns, security risks, verification, and the limitations of AI-generated information.
More importantly, create a culture where employees are expected to question AI. The most dangerous employee isn’t the person who asks, “Are we sure this is right?” It’s the person who assumes the machine must know better.
Critical thinking has become an AI risk-management skill.
4. Measure the Human Impact
Stop measuring AI success only by hours saved, productivity gained, or dollars reduced. Measure what happens to workload, trust, job quality, psychological safety, communication, customer experience, decision quality, and employee wellbeing.
If AI saves 1,000 hours of labor while destroying employee trust, that’s not successful implementation. Efficiency is one metric. Human impact is another. Leaders are responsible for both.
Don’t Automate Away the Skills We Need Most
The more powerful technology becomes, the more valuable deep human leadership skills become.
- Critical thinking
- Empathy
- Courage
- Accountability
- Communication
- Judgment
- Curiosity
- The ability to disagree constructively
- The ability to recognize when something technically possible isn’t ethically responsible
AI can produce reports, analyze data, write emails, summarize meetings, and generate strategies. That makes human judgment more important, not less. Organizations need to develop these capabilities aggressively alongside AI adoption.
Our training programs help organizations strengthen the human skills AI can’t replace. Every program is tailored to the organization’s culture, challenges, and goals, so employees aren’t just learning concepts; they’re building skills they can actually use at work. Check out our menu of training topics here.
Automating tasks makes sense. Automating judgment does not.
Don’t Wait for Damage to Prove the Warning Was Valid
Nobody knows exactly what AI will be capable of ten years from now. Leaders don’t need that answer to make better decisions today. Most importantly, don’t allow the pursuit of innovation to outrun responsibility.
I’ve spent my career watching organizations learn this lesson the hard way: Problems are much easier to prevent than they are to repair.
Toxic workplaces don’t become toxic overnight. Small behaviors get tolerated. Exceptions become habits. Habits become norms. Employees stop speaking up because speaking up doesn’t change anything.
Then leadership looks around and asks: How did we get here?
AI risk develops the same way. Not through one dramatic decision, but through thousands of small decisions nobody questions because everyone is moving too fast.
AI is changing work fast. Make sure your leaders and employees are ready for it. Talk with our workplace culture expert about tailored training that strengthens communication, accountability, critical thinking, psychological safety, and leadership. Let’s build the skills your workforce needs for what’s next.
PS: Do you want to share your expertise with the Civility Partners community? We’re looking for experts to guest our Workplace Intelligence Series’ bi-monthly webinars and Catherine’s LinkedIn Live conversations. If you have insights our audience of HR professionals and leaders needs to hear, we’d love to connect. Fill out the form, and we’ll send you more information about guest opportunities!
