Why Is Psychological Safety Collapsing on AI Teams? | AI for executives

Send us Fan Mail

The neuroscience of why AI pressure in a leader's nervous system destroys team trust before a single word is spoken.

Psychological safety is collapsing on AI teams right now. And most leaders are trying to fix it with better communication. That is the wrong tool for this problem.

 Sahar Andrade, MB.BCh, Neuroleadership Coach and Forbes Coaches Council member, breaks down the neuroscience of what is actually happening. When a leader carries AI uncertainty in their nervous system, their team absorbs it through mirror neurons before a single word is spoken. Cortisol spreads. Threat perception rises. People stop speaking up.

 This is not a culture problem. It is a biology problem.

In this episode: why BCG's 2026 data shows 41% of AI users report increased cognitive load, why HBR found middle managers hold the lowest psychological safety of any organizational layer, what co-regulation actually means and why it matters more than your communication strategy, and what regulated leadership does for team safety at the nervous system level.

The solution is not a new leadership program. It is a regulated nervous system. That is what the B.R.A.I.N.™ framework builds.

Book a Leadership Clarity Call: calendly.com/saharandrade

#PsychologicalSafety #AILeadership #NeuroleadershipCoach #ExecutiveLeadership #NeuroscienceLeadership #AIIntegration #MiddleManagerBurnout #TeamTrust #LeadershipDevelopment #AIForExecutives  #VulnerableLeadership #AuthenticLeadership #AIIntegration #TrustBuilding #Neuroleadership #humancenteredAI #humancenteredleadership  #AIStrategy  #AITransformation #RegulatedLeadership #neuroscienceinleadership #Futureofwork #AIpodcast #AICaféConversations #neuroscience #AItoolsforexecutives #NeuroscienceOfLeadership #LeadershipDevelopment  #aistrategyforexecutives #AIleadershiptransformation #nervoussystemregulation #Neurosciencebasedleadership #AICulturechange #peoplefirst #neuroscienceinAI #AIintegration #AIadoption  

1. Why is psychological safety collapsing during AI rollout?

2. What does a leader's nervous system have to do with team trust?

3. Why do teams lose trust even when leaders say the right things about AI?

4. How does AI pressure in a leader's body affect their team?

5. What is the neuroscience of psychological safety at work?

6. How can leaders rebuild psychological safety on AI teams?

Support the show

--- 

AI Cafe Conversations: Neuroscience-based AI leadership for executives. Hosted by Sahar (The AI Whisperer) | New episodes Wed & Fri 

🔗 Connect: https://www.linkedin.com/in/saharandradespeaker/

📧 Work with me: [e-mail]

🌐 Website: https://www.saharconsulting.com/

 📧 Instagram: https://www.instagram.com/saharthereinventcoach

Episode Transcript

SPEAKER_00 (00:12):
Something happened in a session I had with a
leadership team six months intotheir AI rollout.
The executive in charge said allthe right things.
She told her team that AI was anopportunity, not a threat.

(00:32):
She said nobody would lose theirjob.
She created open door meetingsfor questions.
She did everything theleadership playbook says to the
and her team stopped talking toher.
Not loudly, not dramatically.

(00:54):
They just went quiet.
Fewer questions, fewer ideas,shorter meetings, polished
answers with nothing realunderneath them.
She came to me confused.
She said, I do not understandwhat I'm doing wrong.

(01:16):
I am being transparent.
I am being available.
I'm doing everything right.
And I asked her one question Howdo you feel when you walk into
those AI meetings?
She paused for a long time.
Then she said, Honestly, likeI'm holding my breath the whole

(01:41):
time.
Like I need to look like I havethis figured out, even though I
genuinely do not know if any ofthis is going to work.
And there it is.
That is what we're talking abouttoday.
You're listening to AI CafeConversations, I'm Sahar

(02:04):
Andradi, Neuroleadership Coachand Force Coach Council member.
And today we are going intosomething that is destroying
team trust on AI teams rightnow.
And if you're wondering, thisaccount or this client that I'm
talking about was apharmaceutical company.

(02:25):
I had some comments that came tome asking why I don't share the
name of the executives and thename of the accounts that I work
with on the podcast.
And actually, I really cannot dothat.
We sign agreements, uh, NDAs.
I can talk about the case as acase study, but I cannot mention
any clients by name, either theorganization or the executive

(02:49):
I'm working with for privacypurposes and laws and privacy
laws that protect all myclients.
So I want to make that veryclear.
I will never mention a name ofan organization or a name of an
executive, but I will usewhatever I do as case studies in
anonymous way.
So I just wanted to bring thatto the surface.

(03:11):
But they're all real cases thatactually happen during my work
with executives andorganizations.
Okay.
Going back to what we weredoing, what we were talking
today is what is destroying teamtrust on AI teams right now.
And it has nothing to do withyour communication strategy, it

(03:32):
has everything to do with yournervous system.
So what is the crisis?
What is actually happening topsychological safety right now?
Let me give you some numbersfirst.
BCG, 2026 AI at work survey,found that 41% of regular AI

(03:55):
users report increased cognitiveload.
Not decreased, increased.
Sharm found that 61% of workerstrust senior leaders when it
comes to AI conversations.

(04:16):
But that number drops to 47%among individual contributors.
That was SHARM as HRL.
That gap between what leadersthink they are communicating and
what teams actually feel, thatis not a messaging problem.
That is a nervous systemproblem.

(04:38):
And the HBR, the HarvardBusiness Review research that
came out in October of 2025 madeit even clearer.
Middle managers, the peopledirectly between Executive AI
strategy and frontline teamanxiety, report the lowest
psychological safety of anylayer in the organization.

(05:01):
The lowest.
Not the most junior employees,not the most senior leaders, the
people in the middle.
The ones responsible fortranslating AI decisions
downward, managing team, fear,hitting productivity targets and
still showing up steady.

(05:22):
Think about what that means.
The person your team looks tofor safety during an AI rollout
is the person who feels theleast safe.
And here is the part nobody'stalking about.
The nervous system isbroadcasting that.

(06:45):
This is why emotional contagionis not a metaphor, it's a
measurable neurological event.
Research shows that a leader'scortisol levels, the stress
hormone, are measurable in theirteam members within minutes of
shared interaction.
You cannot hide your nervoussystem state from people who

(07:10):
spend significant time aroundyou.
You can hide your words, youcannot hide your biology.
Now add AI to the combination.
A leader walks into a meetingabout an AI implementation they
do not fully trust.
They feel the pressure of beingaccountable for outputs that

(07:33):
cannot fully verify.
They are aware at some levelwhat their team is watching them
for signals about whether it'ssafe.
Their amygdala, the brain'sthreat detection system
registers this as a complexsocial threat.

(07:54):
The stress response activates,cortisol rises.
The prefrontal cortex, where noone's judgment and empathy live,
starts going offline.
Caput.
And the leader tries to lookcalm while their nervous system
is in full scanning mode.

(08:16):
The team feels it.
Their amygdalas register it.
And they make a rapid andconscious calculation.
If the leader does not feel safehere, this is not a safe place.
Psychological safety does notcollapse because of bad

(08:36):
communication.
It collapses because ofunregulated nervous systems.
And here is what makes thisspecific to AI.
BCG found that leaders are nowspending more time managing and
directing AI than doing theactual work itself.

(08:57):
Nearly half of respondents.
The burden has shifted fromproduction to supervision.
Leaders are in continuousverification mode.
Checking outputs, reviewingresults, staying accountable for
things they cannot fullycontrol.

(09:20):
The nervous system reads this asthreat is ongoing.
Stay alert, do not relax.
That chronic low-gradeactivation never gets discharged
in a meeting.
It just transfers.

(09:41):
Team to team, layer by layer.
This is what I call a nervoussystem cascade.
And it's happening silently inorganizations everywhere right
now.
Let me tell you what it lookslike on the ground.
A director I worked with wasleading a team of 12 through an

(10:07):
AI process overhaul.
She was competent, genuinelyskilled.
She has navigated hardtransitions before, but this one
felt different to her.
She told me, I feel like I haveto look like I know where this
is going, even though thereality is the goalposts keep

(10:32):
moving, and I'm figuring thisout as I go.
Within three months, two of herstrongest performers had gone
white.
They stopped bringing ideas.
They started double-checkingeverything before they shared
it.
They became cautious in a waythey had never been before.

(10:57):
She thought they weredisengaged.
Actually, they were notdisengaged.
They were reading her.

(11:34):
So what is the solution?
Okay, I always want to make surethat we have a solution or a way
out or at least a beginning of aconversation.
So if you want to continue theconversation with me, either
comment down on the podcast orsend me an email at sohar at
soharconsulting.com.
I would love to engage in aconversation with you.

(11:55):
Either you agree or disagree, wecan have a good conversation
about that.
And if you want to be a guest onthe podcast, please let me know.
Our podcast is one of the top 2%local podcasts, right?
Um I just want to make sure thatthe guests speak about AI,
human-centered AI, and how itaffects human beings.

(12:19):
Not in a technological way, justin a human way.
I would love to have you as aguest.
Just email me again at Sahar atsaharconsulting.com.
So let's look at the solution.
What regulated leadershipactually does for team safety.
Here is what the research shows.
Psychological safety increasesteam learning behavior by 35%

(12:42):
and team performance by 27%.
That is Google's projectAristotle data.
I'm just not giving you figuresthat I just made out.
This, whatever research I share,comes from credible sources.
So psychological safetyincreases team learning behavior

(13:03):
by 35% and team performance by27%.
Again, Google Project Aristotledata.
But here is what ProjectAristotle did not fully capture.
Psychological safety is notbuilt through ground rules and
speaking norms.
It's built throughco-regulation.

(13:24):
Co-regulation is the biologicalprocess through which one
nervous system helps anothercalm down.
It happens through tone, throughpace, through facial expression,
through the quality of presencesomeone brings into a room.

(13:45):
When your nervous system isregulated, the people around you
feel it.
Not metaphorically,neurologically.
Research on the polyvagalsystem, the part of your nervous
system that governs socialconnection and safety, shows
that humans are constantlyreading environmental signals

(14:06):
for cues about threat andsafety.
The primary signal source in anyteam environment, did you guess
what it is?
It's the leader.
You are a thermostat, not athermometer.
A thermometer beats thetemperature of the room.

(14:26):
A thermostat sets it.
Every leader is setting theemotional temperature of their
team.
The question is whether you'redoing it consciously or by
default.
Here is what changes when aleader regulates first.
The team's threat detectionsystem receives a totally

(14:50):
different signal.
Not danger is here.
The leader is not panicking.
This must be manageable.
This is what a team perceivesand feels.
So their prefrontal cortex staysonline.
Their creative thinking staysavailable.

(15:11):
People speak up instead ofpulling back.
Ideas move and flow instead ofgetting suppressed.
That is psychological sympathy.
Not a cultural program, abiology outcome.
Now I want to be very clearabout something.

(15:31):
This is not about performingcalm or, you know, woo-hoo out
there and sit on the floor foran hour for meditation.
Though meditation is great, butthat's not what I'm talking
about.
This is about the differencebetween suppressing your nervous
system response and actuallyregulating it.
Suppression looks calm.

(15:54):
The team still feels itunderneath every single time.
Regulation is different.
Regulation means your nervoussystem has actually moved out of
threat mode and into a statewhere you can think, connect,

(16:14):
and lead.
The body does not lie, it justgets very good at broadcasting
the truth before you don't.
In my work with leaders, I use aframework called Brain
V-R-A-I-N.

(16:35):
I'm not going to walk youthrough all the five parts
today, but I will tell you aboutthe first one because it's the
one most leaders skip straightpast.
The first piece is not acommunication tactic.
It's not a mindset shift.
It's a physical recept that hasto happen in the body before any

(17:01):
of the rest of it works.
Most leaders try to fix teamtrust by saying better things.
Closing and choosing the rightwords.
Getting the messaging right.
But your nervous system does notwait for you to find the right
words.
It's already broadcasting beforeyou even open your mouth.

(17:24):
So the first move is not aboutwhat you say, it's about what
state you are in before you sayanything at all.
That one shift, just that firstpiece, changes how an entire AI
rollout conversation goes.

(17:46):
Not because the leader gotsmarter, because the leader got
regulated first.
So the B is actually forbehavior.
I will not unpack the other fourhere.
I will do it subsequently as wego through the podcast.
But that first one alone isworth sitting with.

(20:18):
So let me push back at thatbecause it misses what I'm
actually saying.
I am not telling you to feelcalm when you're not just like,
yeah, go zen and you that bymiracle you're gonna feel calm.
That's not what I'm saying.
I'm not telling you to fakeserenity in front of your team
either.
And I'm not subscribed to fakeit till you make it.

(20:42):
I'm not there.
Both of them are performance.
And performance is exactly whatyour team's nervous system can
detect and reject.
What I'm telling you here isdifferent.
The leaders who holdpsychological safety together
during chaos are not the oneswho have no uncertainty.

(21:05):
They are the ones who havelearned to be honest about their
uncertainty from a regulatedstate instead of a panicked one.
That is a skill, my friends, andthat is trainable.
It's not about fewer hardmeetings, it's about walking
into the hard meeting with anervous system that can stay

(21:29):
present instead of one that issilently broadcasting alarm.
That is one more piece.
And I'm gonna add another one.
The leaders I see maintainingpsychological safety through AI
transitions are not the oneswith the best strategy text.

(21:52):
They are the ones who do threethings consistently.
First, they regulate before theylead.
Do not walk into high-stakes AIconversations, still carrying
the anxiety from the lastmeeting.
They create intentionaltransition.
A walk, a breath, a pause beforethe door opens.

(22:13):
And again, I got pushed back onthat.
Well, if it was that easy, well,it's actually that easy.
What you're doing, you areinterrupting a cycle.
By going for a walk, a breath,or a pause before the door
opens, you're giving your brainthe space to reset.
And our human brain physiologyand anatomy are very easily

(22:37):
distracted.
It's as easy as pushing a resetbutton on a laptop.
It resets itself.
It has to become a habit, atrigger that we create so we can
reset our nervous system beforewe walk into a meeting.
Second, they make it safe to notknow.
They model uncertainty asinformation, not failure.

(23:00):
They say it out loud.
I don't have this figured outcompletely.
Here is what I know.
Here is what we are stilllearning.
And what I hear from a lot of myclients is that I don't want to
appear weak.
I don't want to appear like Idon't know what I'm doing.
So I'm not going to ask for helpand I'm not going to say I don't
know.
And this is being vulnerable isactually your biggest strength.

(23:24):
Saying, let me figure it outbecause I still don't know, and
I'm figuring it out with you iswhat actually creates the
foundation of trust andpsychological safety because you
are truthful.
Third, they stay curious insteadof certain.
In ambiguous situations,certainty performed is a threat

(23:45):
signal.
Genuine curiosity is a safetysignal.
The brain registers thedifference every single time.
None of this requires a newleadership program.
It requires a regulated nervoussystem, which is exactly what my
brain framework is decideddesigned to build.

(24:09):
Not thought process, not throughinsight, through practice.
So let me bring this.
They're not words on a websiteor an MOU or on a guideline or

(24:33):
even an ROE.
It's collapsing because leadersare carrying AI uncertainty in
their nervous systems and theirteams are absorbing it.
Mirror neurons do not read yourslide deck, they read your body.
You cannot train your weight outof a nervous system state.
You have to regulate your weightout of it.

(24:54):
When you do, something shiftsfor your team.
Not because you said somethingdifferent, because you became
something different in the room.
And the room felt it.
That is the work.
If you're leading a team throughan AI transition and you
recognize what we talked abouttoday, I want to have a

(25:14):
conversation with you.
If you want to be a guest on mypodcast, email me.
30 minutes.
No pitch, just clarity where youactually stand and what your
next move can be.
The link for the book and theshow notes.
Thank you for being here.
Share this with the leader whois trying to hold it together

(25:35):
right now.
They need to hear that whattheir team is feeling is not a
trust problem, it's a biologyproblem.
And biology has a solution.
I will see you on Friday on thearticle, uh, Forbes article like
short form of our podcast.
Before I leave, as I always say,show me some love.

(25:57):
Comment, rate, subscribe, shareour podcast with someone else.
Thank you for supporting us andmaking us one of the top 2%
global podcast out there.
I love you and thank you foryour support.
Till we meet again, come up,come up, come on.