This week, three very different things pointed in the same direction.
Brazil’s NR-1 regulation came into full enforcement, placing psychosocial risks: burnout, harassment, work overload, on the same legal level as physical accident risk. The IDC published a study showing that companies invest up to 14 times more in AI tools than in developing the people who will use them. And a major tech player published on their corporate blog a headline that stopped me cold: “AI needs more than intelligence, it needs humanity.”
This is not a coincidence. It is a signal.
NR-1 acknowledges what we already knew: burnout and overload are not solved by internal engagement campaigns, meditation apps, free snacks, or the occasional day off. These risks are created by company culture and by the way leaders behave.
The IDC puts numbers to it: companies allocate between $6.5M and $14.4M in AI tools — and less than $1M in leadership training. The result: fewer than 1% of organizations have reached the ideal stage of human development for the AI era.
The report identifies building psychological safety as an indispensable leadership skill for AI to actually work because only in a safe environment will employees experiment, fail, and learn with new technology.
Technology racing ahead. Humanity falling behind.
The study lists five categories of human skills that no AI can replace. The one that interests me most because it has been my focus and research for years, is humanized leadership: empathy, change management, and psychological safety.
Compassion is not just empathy
This distinction matters and it is where most leadership programs fall short.
Empathy is feeling what the other person feels. Compassion is feeling and acting. An empathetic leader absorbs the team’s pain. A compassionate leader transforms that perception into something concrete: the right question, the safe space, the decision that protects the person before protecting the process.
Neuroscience proves this: when a leader is genuinely present, the employee’s brain exits threat mode, where creativity and learning are blocked, and enters connection mode, where real performance happens. This is not life-coach talk. It is physiology.
And this is exactly the kind of safety that AI adoption demands. To test a new tool, make a mistake in front of a manager, propose a different workflow, or admit they didn’t understand the model — the person needs to feel safe. That safety does not come from the platform. It comes from the leader.
Ever since I entered the corporate world, regardless of company or industry, I had to navigate the famous quarterly reviews — meetings where teams presented results in polished, well-rehearsed decks. Mistakes carefully swept under the rug. Sometimes at the cost of wasted budget, hidden rework, burnout, and lessons that never saw the light of day.
I decided to change the format. Instead of only showing what worked, I created what I called a lowlights deck: what didn’t go as planned, what could have been done better, or cancelled altogether. Each team shared one thing that hadn’t worked and what they would do differently if they could go back. It was light, sometimes even fun, results often came with laughter, but with real substance behind it: accountability without punishment, learning without humiliation.
The effect was immediate. People started showing up more honest, more open, more willing to try new things because they knew mistakes would not be used against them. The quality of decisions improved because real problems could finally be named.
That is compassion applied to leadership. And it is a skill we can learn and practice.
How am I creating the conditions for the people around me to feel safe enough to try, fail, and learn?
NR-1 sets the legal floor. The IDC data shows the gap. But the ceiling, what separates an environment that merely complies with the law from one where people truly flourish, is built by leaders who chose to put humanity at the center of their work.
I wrote The Power of Compassion in the Workplace exactly for those leaders. With updated data, practical frameworks, and everything I learned in over 25 years inside large organizations.
The book is available on Amazon Brasil and you also can visit my site http://lisapolloni.com
Sources: IDC InfoBrief “Powering Up: Human Skills for the AI Era” (May 2026) | Microsoft Cloud Blog “AI needs more than intelligence, it needs humanity” (May 2026) | Brazilian Ministry of Labor: NR-1 updated regulation (punitive enforcement: May 26, 2026).
#CompassionateLeadership #ArtificialIntelligence #Compassion #PsychologicalSafety #FutureOfWork #HR #PeopleManagement #Burnout #HumanizedLeadership #WorkplaceWellbeing
