friends ≠ free passes
Hi ya party people,
Happy Black Friday, y’all! I hope your Thanksgiving was full of carbs, naps, and zero conversations about how “AI is going to replace HR anyway.”
If you’re reading this while hiding from a mountain of leftovers or ignoring a family member’s take on “what HR actually does,” just know…I see you, and I’m thankful for you. 👀
So let’s get right into our Q&A this week: how to deploy AI in HR without accidentally coding chaos into your org, and what to do when certain coworkers think friendship is a disciplinary force field.
ALSO, I’m declaring Dec. 5th a Safe Space Holiday because we have over a dozen IRL Meetups taking place to celebrate the last First Friday of the month! 🥳
Here’s what’s already on the cal: Austin, Atlanta, Baltimore, Charlotte, Chicago, Dallas, Denver, Houston, Las Vegas, Los Angeles, Madison, Minneapolis, Salt Lake City, Seattle, Scottsdale, SF/Bay Area, NoVa/DMV, and Virginia Beach.
Don’t see your city? Submit an event!
And keep up the question submissions!
ASK AWAY
🤔 coworkers & biasesCopy anchor linkCopied
✍🏾 How can organizations deploying AI in HR ensure outputs are fair, accurate, and free of bias and/or hallucinations?
Context: Tech company, approx. 1000pp. Tech side of business is deep in AI, corporate functions are slow adopters and ready to get onboard, with proper guardrails.
📣 Nalini Mani, Organization Transformation Advisor @ N/A:
To ensure AI in HR is fair, accurate, and free of bias or hallucinations, organizations must combine technical rigor with human governance. That means curating representative data, standardizing inputs and outputs, and applying bias‑mitigation techniques.
Human oversight remains essential, supported by explainability tools and clear audit trails. Just as important, HR teams need AI literacy to interpret outputs responsibly, and partnering with I/O psychologists ensures systems align with validated principles of fairness, assessment, and employee well‑being. Together, these guardrails create trust while enabling responsible innovation. To be more detailed:
Data integrity & fairness: Models must be trained and tested on representative datasets, with bias-mitigation techniques applied and irrelevant identifiers (like names or schools) stripped from inputs.
Input/output hygiene: Standardized, structured inputs reduce ambiguity, while output gating and confidence thresholds prevent low-quality or hallucinated results from reaching HR teams.
Human oversight: AI should augment..not replace…HR judgment. Sensitive decisions like hiring or promotions must remain human-in-the-loop, with clear audit trails for accountability.
Transparency & explainability: HR teams need interpretable outputs, not black-box recommendations. Explainability tools help build trust and allow decisions to be challenged.
Tech literacy & adoption: Corporate functions must be upskilled to understand AI’s strengths and limitations. Building AI literacy ensures HR teams can spot bias, question outputs, and escalate concerns rather than blindly relying on automation.
I/O psychology integration: Partnering with industrial-organizational psychologists ensures AI systems align with validated principles of assessment, fairness, and employee well-being. Their expertise helps design evaluation frameworks that go beyond technical accuracy to consider human behavior, motivation, and organizational culture.
Governance & culture: An ethics committee spanning tech, HR, legal, DEI, and I/O psychology should oversee deployments, while employees are informed about when and how AI is used. Clear communication and opt-out mechanisms build trust.
In short, fairness and accuracy come from combining robust technical design, disciplined input/output practices, human governance, and behavioral science expertise. That way, AI in HR avoids the pitfalls of bias or hallucination.
📣 Rebecca Taylor, Co-Founder/CCO @ SkillCycle:
The truth is, you can't guarantee that the output is free of bias or hallucinations - just like you can't believe everything you read on the internet. But you can build your AI literacy and train yourself and employees to fact-check, recognize bias, and prompt properly.
And +1 to Nalini's points about standardizing inputs and outputs, and curating representative data. AI literacy is the most important soft skill we all need to be building right now as the tech continues to become more embedded into our workflows.
📣 Marya Calhoun, SVP HR @ Affordable Housing:
The real guardrails come from process, not the tech. Use a human-in-the-loop approach so HR reviews outputs for fairness, accuracy, and context. Put a simple governance model in place: routine bias checks, audit trails, and quick escalation steps when something feels off.
Roll out AI in phases. Pilot first, measure results, refine, then scale. When people understand how the tools work and have clear checkpoints, you cut down hallucinations and build trust across the organization.
Safe Space members can join this discussion here. Not a member yet? Apply to join here.
✍🏾 How to deal with co-workers who think they shouldn't be reprimanded because you get along with them? They think they are exempt.
Context: 58 staff
📣 Elizabeth Lee, HR Consultant @ LIV Golf:
There's business relationships and personal relationships all within the same employee. As a P&C professional, it's key to start off conversations with those you have strong personal relationships with, and make it clear that this is a discussion about work, and focus the conversation on the situation, the impact of their behavior to the business and the expected next steps or repercussions of one's actions - and document the conversation.
You can still be human and empathetic, though not to the point of undermining your message to the colleague about the issue at hand. Don't shy away from tough conversations with work friends - you're not doing anyone any favors by doing so.
📣 Alex White, Regional Director, HR @ Washington Duke Inn & Golf Club:
This is such a real challenge — and one that more leaders face than people realize. I would like to approach this from both a professional and very personal place:
At its core, this is about boundaries. When people start believing that a good relationship equals immunity, it usually means expectations haven’t been reinforced clearly — or consistently — enough.
And I say that with full understanding, because years ago I learned this lesson the hardest way possible. I had to terminate the employment of someone who was dating one of my closest friends. He told me, point blank, that if I went through with it, he and I were done.
I remember telling him as calmly as I could:
“If you can’t separate business from personal, then the friendship I thought we had isn’t what I believed it to be.”
And unfortunately, that was the last conversation we ever had.
It was painful — but it taught me something I’ve carried with me ever since:
You cannot lead effectively if people believe your personal relationships override accountability. You can care deeply about people and still hold them to expectations. In fact, that’s what real leadership requires.
A few things that help in situations like this:
- Be consistent. When everyone sees that the same standards apply to all, the “but we’re cool, right?” objections disappear.
- Be clear. Remind them kindly but firmly: the relationship doesn’t change the responsibility.
- Be steady. The more calm, fair, and predictable you are, the less room there is for people to push boundaries.
And if someone “gets offended” that they were held accountable? That usually tells you everything you need to know about whether their respect was genuine or conditional.
At the end of the day, professionalism isn’t about distance — it’s about integrity. And integrity means treating people well… while still doing the job you’re entrusted to do.
📣 Alexandra E, CHRO @ Credit Union:
I don’t have context for the OP’s relationship with the person they’re using as the example; not sure if they were peers previously and now OP is a supervisor? Regardless, hopefully this helps:
You treat them kindly and fairly—but not specially.
This is basically a boundaries + expectations problem.
- Get your mindset straight first
Being friendly is not equal to being exempt.
Your job is to be fair and consistent, not to protect people from discomfort.
If you don’t hold them accountable, you’re actually being unfair to the rest of the team (and to them, long term).
Remind yourself: you’re not betraying them; you’re doing your job. - Be very clear on the behavior, not the relationship:
Before you talk to them, jot down:
What exactly is the issue? (late, misses deadlines, attitude, not following process, etc.)
Dates / examples
What’s the impact? (More work on others? Slows the process? Hurts members/customers?)
That keeps the conversation grounded:
“This happened, here’s how it affects the team, here’s what needs to change.” - Have the conversation directly (but not dramatically)
Do this 1:1, in private. You can acknowledge the relationship briefly, but then pivot to expectations.
“I want to talk about something important.
We get along well, and I value that. But I need to be really clear: my responsibility is to hold everyone to the same standards, including people I’m close to.
Lately I’ve noticed [behavior: e.g., frequent late arrivals / missed deadlines / not following X process] on [specific dates/examples].
This causes [impact – delays, extra workload, member experience, etc.]. Going forward, I need you to [specific change].
If it doesn’t improve, I’ll have to move to formal documentation, just like I would with anyone else.”
If they say something like, “But I thought we were good / I thought we were cool”:
“We are good. This isn’t about liking you or not liking you. It’s about the role I’m in and the standards for this job. I can like you and still need you to meet expectations.” - Don’t over-explain or apologize for holding a boundary:
Avoid:
“I’m so sorry I have to do this…”
“I really don’t want to write you up but…”
Instead, calm and factual:
“This is the next step in our process.”
“I’d rather correct it now than let it hurt your record later.”
“This is the same way I handle it with everyone.”
Friendly tone, firm content. - Follow through consistently
People believe what you do, not what you say.
Document the conversation (even just notes to yourself).
If the behavior continues, follow your normal process: verbal : written : whatever your HR steps are.
Apply the same standard to others so it doesn’t look personal or targeted. So if others are doing the same things, they get the same disciplinary actions.
If they test the boundary (“you’re really going to write me up?”):
“Yes. This is the same step I’d take with anyone else in this situation. I’d rather it be predictable and fair for everyone than arbitrary. - Protect yourself
Because this kind of person will sometimes flip the story when they don’t get special treatment, keep emails/notes on performance and expectations.
After the 1:1, you can send a brief recap email:
“Thanks for meeting today. As discussed, the expectation is [X], and we’ll check back on [date] to see how it’s going.”
If you have HR, loop them in early—not as a “hammer” but to support your documentation and provide backup.
Hope that helps. Good luck!
Safe Space members can join this discussion here. Not a member yet? Apply to join here.
🚨 ON YOUR RADAR
🎧 This week’s episode is chaotic in the best way because I sat down and chatted with my actual direct report, Madison Montgomery! We unpacked generational gaps, the infamous “generational shits” typo, and what happens when Millennial HR meets Gen Z energy. Check it out on Spotify or Apple Podcasts!
💰 Pay Transparency matters (but it's also difficult). That's why I created this course. Join the Safe Space community to access it, plus everything else you need to stay ahead of the curve.
📝 RESOURCE OF THE WEEK
Each week I feature a resource I love from the Safe Space library that I believe would be relatable to this week's newsletter topic.
This week, we’re talking about appreciation in the workplace. Check it out HERE ⬇️
FRIDAY FUN
😂 me 24/7 while also WFHCopy anchor linkCopied
_j7m6.png)
With holidays coming up, is this you??? :)
Does anyone remember the Sabrina the Teenage Witch episode with Britney????
I SURE DO!!!!! If you need me, I'll be rewatching this today.
That's all for this week! I hope you enjoyed! If you have any thoughts, please let me know. I'm allll ears.
Reminder: Today is FRIDAY. 🙏
