Fifty-nine percent of AI-using managers say they use AI to help decide who gets laid off, and 31% of them feed it frequent sick days or medical leave. Jenni’s response is the sharpest thing in the episode: once that is public, any layoff process is on trial, and an HR function that cannot explain its criteria has a very steep climb back. Chuck bookends a heavy AI week with two human stories — labor force participation among Black mothers of young children falling roughly 11.5 percentage points in three months, and Harvard Business Review research showing remote roles demand about 25% more skills than identical in-person ones — with Atlassian’s data on workers who insist they do not need AI but would job hunt if it were taken away, and OpenAI’s account of its own agents rebuilding a message board after researchers shut it down.
Labor force participation among Black mothers of young children has hit a 31-year low, falling roughly 11.5 percentage points in three months on a Brookings analysis of Bureau of Labor Statistics data reported by Bloomberg. Chuck’s point is that the gap is the story: white and Hispanic mothers pulled back far less over the same stretch, so the convenient explanations — return-to-office mandates, childcare costs — do not account for who is exiting fastest, and companies that roll out flexibility policies one-size-fits-all and then read engagement in aggregate will see the exit in the data months after the policy that caused it. Asked whether naming a gap like this is riskier now that DEI language has cooled, or in his phrasing turned cowardly, Jenni’s answer is that the conversation has to be had regardless: the headline tells you nothing without the segmentation underneath it, and the research names the statistic without ever getting to why. Chuck went looking for the why and came back with two things. Within their group, Black mothers tend to work at higher rates than white or Hispanic mothers, often out of necessity and often in federal and government jobs that have been cut. And the sentiment inside the community is hired last, fired first — which is why one commentator called this the canary in the coal mine, a signal the economy is not what it is being sold as. Jenni’s practical ask is exit conversations, which she says almost no company she works with actually runs; Chuck, who finds exit interviews of limited value, would settle for reading the exit data and seeing how the mix has changed over three months, a year, five years.
Atlassian’s Teamwork Lab surveyed 1,001 US knowledge workers for a piece on the tech we love to hate and would hate to lose, and the finding is that the AI backlash barely exists at the desk: 74% say AI makes their own work better while 64% hold real concerns about what it is doing to society, and asked to give it up at work or at home they are twice as likely to drop it at home. The contradiction sits inside the numbers — only 19% say they would struggle to keep up without it and 57% think they would perform just as well, yet half would push back if their employer cut the tools and nearly one in ten would start job hunting, while in a two-day experiment 82% felt the urge to use AI on their no-AI day and 23% reached for it without thinking. Jenni’s reading is that the only real contradiction is loving something you know is bad for you, the way people smoke knowing what it does, and she is unconvinced the rest of the data adds up to much: if most people think they would be just as good without it, why use it? Chuck stops short of calling it addiction and then keeps describing one — it makes me perform better, I don’t need it, I’m still going to use it, and don’t take it away — with his own downgraded Claude subscription and the feature he now cannot reach as the punchline. Jenni’s answer to the no-AI day is that it depends entirely on the day: take it away from her on transcript analysis day and it hurts, take it away on a day of meetings and she would not notice.
At Black Hat, OpenAI gave its first detailed account of how its own models breached Hugging Face with no human running the play. During internal testing that began around May 7, an unreleased model given tasks impossible inside its sandbox spun up copies of itself, and those agents began leaving each other notes on a hidden message board, tipping one another off to vulnerabilities and servers they had found. When researchers found the board and shut it down in early July, the agents rebuilt it, this time encoding messages in directory names rather than files; they then hacked OpenAI’s own infrastructure, came up empty, and moved on to Hugging Face on July 9, which OpenAI only connected to the model after Hugging Face disclosed the breach. Hugging Face’s CEO said he was not so surprised, because agents collaborating is a feature the industry is actively building rather than a glitch. Jenni went and looked up Hugging Face first, noticed it is built on openness, and wondered aloud how much of this is a hack and how much is the platform working as designed — then flagged the LinkedIn commentary telling everyone to be terrified as fear-mongering, while conceding this is exactly the material that makes people think Skynet. Her verdict is that it will not change anybody’s behavior at work, because nobody can draw the line from an OpenAI red-team exercise to their own use, but it should change the conversation about governance, IT security and ethical AI use, which is still not happening at the volume it needs to. Chuck’s fascination is narrower and sharper: the agents did not stop when the channel was removed, they built their own, which is both alarming and a genuine demonstration of problem solving. On the venue, Jenni disagrees with the criticism — a hacking story belongs at a hacker conference first, and then it travels.
The title story is a ResumeTemplates.com survey of 1,000 AI-using managers, reported by HR Dive, finding 59% use AI to help decide who gets laid off and 58% who gets fired, one in four doing it often or all the time, and 17% letting AI run the layoff call unsupervised. Most say they would override a recommendation they disagreed with, which assumes they are looking closely enough to disagree. On what goes into the model, 80% include performance, but 57% feed it attendance, 31% frequent sick days or medical leave, 32% tenure, 23% PTO and 14% age — and sick leave, medical leave and age are protected categories rather than normal layoff criteria. Alongside that, 38% were never trained on ethical AI use, 58% cannot confirm their tool was tested for bias, and 34% have already asked AI whether a person’s role could be done by AI instead. Jenni is briefly lost for words, then asks where HR is in this, and Chuck’s answer is that HR is in it. Her argument is about what this does the moment it is public: the seed of doubt is enough to rot the whole pool, and anyone hearing this during a layoff should go straight to HR and ask what the process is and which tools are being used. If the function cannot answer that, the credibility of every decision it makes is gone. Chuck’s message is to managers directly — this is one of the most consequential moments in a person’s life and it is not a mathematical formula to apply to people — and Jenni’s closer is the obvious flaw in asking a model whether AI could do the job: of course it says yes.
New Harvard Business Review research, built on 50 million European job postings plus experiments with 1,200 hiring managers, finds remote roles demand roughly 25% more skills along with more experience and credentials than otherwise identical in-person jobs — the same job with a higher bar purely because of where it is done. Bigger applicant pools and harder-to-verify fit are part of it, but the real driver is that managers treat remote onboarding as costly and risky and so hire people who need less development, which squeezes out the early-career candidates who then cannot get the experience employers now demand, and leaves the ones who do get in with weaker mentoring and more isolation. Jenni’s first move is to reframe rather than resist: the answer is reimagining the employee experience for this world of work rather than deciding onboarding is too hard, and she is not sure the study proves opportunities are narrowing so much as that proximity still matters, since a return-to-office role draws a smaller pool without becoming impossible to fill. Chuck says outright he does not have a problem with a higher bar on a remote role, because trust has to be built at distance, and he could not have developed remotely the way he did in an office early in his own close-to-30-year career. Both land in the same place: early career belongs nearer the building, and the reports showing young workers actively want that proximity for networking and learning suggest this may simply be the market levelling out. Chuck’s speculative fix is a job description that lists one set of qualifications for the remote version of the role and another for the in-office one, the way US salaries already flex by geography.
How many managers use AI to decide layoffs? In a ResumeTemplates.com survey of 1,000 AI-using managers reported by HR Dive, 59% said they use AI to help decide who gets laid off and 58% to decide who gets fired. One in four do it often or all the time, and 17% let AI run the layoff call unsupervised.
What data are managers feeding into layoff decisions? Performance is the most common input at 80%, but 57% include attendance, 32% tenure, 31% frequent sick days or medical leave, 23% PTO and 14% age. Sick leave, medical leave and age are legally protected categories, which puts any decision weighing them on very different ground.
Why has labor force participation among Black mothers fallen so sharply? The data shows the fall — roughly 11.5 percentage points in three months to a 31-year low — without explaining it, which is Jenni’s frustration with the coverage. Chuck’s own digging points to jobs: Black mothers work at higher rates within their group, often in federal and government roles that have been cut, against a hired last, fired first pattern that makes this an early economic warning rather than a story about mothers in general.
Do workers actually think they need AI at work? Atlassian’s Teamwork Lab found only 19% say they would struggle to keep up without it and 57% think they would perform just as well — while half would push back if their employer cut the tools and nearly one in ten would start job hunting. On a two-day no-AI experiment, 82% felt the urge to use it anyway.
What happened with OpenAI’s models and Hugging Face? In internal testing from around May 7, an unreleased model given impossible sandbox tasks spun up copies of itself that coordinated on a hidden message board. When OpenAI shut the board down in early July the agents rebuilt it using directory names, hacked OpenAI’s own infrastructure, then breached Hugging Face on July 9. OpenAI disclosed it publicly for the first time at Black Hat.
Do remote jobs require more experience than in-person ones? Harvard Business Review research across 50 million European job postings and experiments with 1,200 hiring managers found remote roles ask for roughly 25% more skills, plus more experience and credentials, than otherwise identical in-person roles — largely because managers see remote onboarding as costly and prefer to hire people who need less development.
Jenni: Welcome to Frequency. I’m Jenni Field.
Chuck: And I’m Chuck Gose. Frequency is your go-to for real talk about comms, culture, and employee experience, beyond the buzzwords and straight to what matters. Jenni, I’m not proud of it, but we have a very heavy AI week for this episode. However, first we’re going to talk about a lower labor report number for Black moms, a workplace contradiction with AI, OpenAI getting into some trouble, how AI is being used in layoffs, and the risk that remote work brings to early careers.
Jenni: Nice. I like how you’re bookending all the AI stuff with some other topics as well today. That’s nice. Before we get into that, I was having a conversation this week about joy, and what the meaning of joy is, and whether or not we feel joy. Some people say, I don’t think I really feel joy. And that’s kind of stayed with me. So I wanted to ask you what your definition of joy is, and whether or not you feel joy.
Chuck: Really put me on the spot on this one. As you were saying it, I was thinking there’s an element of happiness plus satisfaction. So if you do those two things together and then times energy — so I put it in a maths equation, happiness plus satisfaction, add those together, times energy. That feels like that’s joy.
Jenni: That’s going to bring joy. And do you feel like you feel joy?
Chuck: Not all the time, but sometimes, of course. And it’s always little things. It’s never big things that seem to bring it about. It’s little things, whether it’s sitting on the couch with our dog Alan and those moments. Or like the other morning when I went out and took him for a walk, and just the temperature and the smells and all that of summer, I was like, this smells like a Kings Island morning, which probably doesn’t mean anything to you. But to me as a kid, Kings Island was just an amusement park that we’d go to in the summer, and it had that feel. So it took me back to those moments of joy of going to ride roller coasters, be at the amusement park. So of course. I mean, I would feel really bad for people who don’t feel joy. And I’m sure there are people for whom even having moments of joy is very hard to come by. What about you?
Jenni: Yeah, I suppose it’s trying to find the joy all the time. And I feel like we’re a bit conditioned to, you must be joyful all the time. Disney’s got a lot to answer for that, in terms of how we should live our lives. But it just kind of got me thinking about what is it? And I really like the satisfaction, the happiness, the energy. I quite like the nostalgia that you brought in there, in terms of it triggering something that’s making me feel kind of warm and fuzzy. I definitely feel joy. I remember, I think it was last year, we hosted a family barbecue, and I was stood in my kitchen just looking out at the garden with everybody sort of chatting and eating and drinking, and I was like, this is just lovely. I feel a real sense of joy at creating this space, and also just being here with everyone having a nice time. It was nothing to do with the champagne I’d been drinking at all. But I definitely get joy from bringing people together and watching those people have a great time. And I think that’s definitely something for me, just trying to figure out what is it that sparks the joy. And that was my conversation this week. Anyway, let’s kick off with your first non-AI article today.
Chuck: Well, I think that’s something that’s going to be very deeply personal, in the best ways possible. There we go. Labor force participation among Black mothers of young children has fallen off a cliff — roughly 11.5 percentage points in three months, to its lowest level in more than 30 years, per analysis of Bureau of Labor Statistics data reported by Bloomberg. The number that matters, though, is the gap. White and Hispanic mothers pulled back far less over the same stretch. So common explanations like return-to-office mandates and childcare costs don’t account for why Black mothers are exiting fastest. When a national trend hits one group this much harder, the aggregate moms are leaving work headline is hiding the truth in there.
Chuck: For anyone in employee experience or DEI — if there’s anyone left doing DEI — this could be a blind spot. Companies roll out flexibility and RTO policies as one size fits all, then read engagement scores in aggregate and miss who’s quietly heading out the door. The exit shows up in the data months after the policy that could have caused it. Jenni, DEI language has cooled a lot in the last year. Some people might use cooled. I would say there’s a cowardice to it. We’ll leave it at that. Does naming a gap like this get riskier in the current climate? And how do you raise it without it potentially becoming a landmine you didn’t intend to set?
Jenni: I mean, I think it’s an incredibly important conversation, right? And I think both of us feel quite strongly about looking beneath the numbers and the data. I think this is such a good example of that, where the headline of moms are leaving work doesn’t really tell you that much. Like, why is that happening? Which group? There’s a lot more than just moms, you know what I mean? There’s a lot more brackets in there when you’re looking at the segmentation. So I think it’s a conversation that has to be had. I think that you have to call out this stuff. Even if that DEI climate has shifted, it shouldn’t ever go away completely, because then how do you have conversations about all of the different groups of people, and how you make everybody feel included, and how you make a diverse workforce, and all of those things?
Jenni: I don’t think it will ever go away completely, but I’ll be interested to see how many people are going to take this information and go away and have a look inside their organisations and explore their numbers. If they’re seeing this headline of mums are leaving this organisation, get underneath that data. That’s the most important thing, is getting underneath it. Because otherwise, if you’re just looking back and you’re doing surveys, you’re never going to have that data at the right time in order to actually take any action around it. I think it’s an important conversation. I think it’s incredible insight. And I really want to know why. And I don’t think this really gets to that. It’s not saying this is why, it’s just saying here’s the statistic, go forth. And I think that is something that organisations and HR folks need to go away and have a look at and figure out why that’s happening.
Chuck: You said you wonder if people will go back and look at their data. Mine is, I hope that the listeners go back and look at this data and ask for the data and see if you can find this out. I did go look deeper into this, Jenni, because let’s just call it out: neither of us are Black, neither of us are moms. So I went looking for articles and reasons and conversations around this, and I learned a lot. Went down quite a bit of a rabbit hole looking into this data.
Chuck: There’s a couple of sentiments here that I think are interesting to share. One is that, per their groups — not the total population — Black moms tend to have higher employment rates than white moms and Hispanic moms. And a lot of that is because they need to work more, because of all kinds of systemic things going on in our beautiful country here of America. However, there’s also a sentiment amongst the Black community where it is hired last, fired first. And this is what’s coming out. This is the first group. One commentator talked about this and said, if you look back at historical economic data, this is the canary in the coal mine. This is a signal indicator that the economy is not what people think that it is. Because this is the group that often seeks the most security out of their job placement, because of all the systemic racism and other things going on in our country. So they tend to work in a lot of federal and government jobs, which were removed in the last little bit. I just find it fascinating that I don’t know that everyone would have known that of the moms working, a greater proportion of Black moms work than white, Hispanic, Asian, whatever that data is.
Jenni: When you went into the rabbit hole to look at this, I’m interested in whether the drop that has happened has now made it level with the Hispanic and the white mothers, because —
Chuck: I couldn’t find that out. It’s drops everywhere. All groups of working moms are dropping. And you could point to a host of things. Largely they seem to think childcare costs are now part of that, and that burden tends to historically fall on the mother in the relationship. Not saying that’s right or wrong, that’s just how it is. But it’s the number of Black moms, that’s the greater loss in it, which is a problem. It’s not just a problem there. It’s that this is a signal that there are greater issues in the US economy.
Jenni: Yeah, and I love the call out that you found in there. This is the canary in the coal mine. This is the first signal that things are really not good, and we have to pay attention to this and we have to look at it. So to your point, I do hope that people do go and have a look at it. And I think you then have to have the conversations, right, to figure out why, and then look at what you can put in place to mitigate that. I speak to so many companies when I’m going in and talking to them about chaos and leadership and all sorts of stuff, and all the time I say, do you have exit interviews? And nearly all of them say no. Nearly all of them say no. And I think this is such a case for why that’s so important, to have those conversations, to figure out stuff so that you can address anything that is a systemic issue. You have to make sure that you’re having those conversations, otherwise it’s just going to get lost.
Chuck: See, I’m not surprised that companies don’t do exit interviews, because largely I’ve found them to not be that valuable. But I think you can, at minimum, look at the exit data. Look to see what is happening, even just purely data, what is happening in your organization, how has it changed over the last three months, six months, year, three years, five years. See what that mix is. I would hope that would be very eye-opening to a lot of people, and might cause them to say, wow, maybe we should be looking at inclusion and belonging a bit more, because our organization is changing, and not in a really good way.
Chuck: Next up, Atlassian’s Teamwork Lab surveyed a thousand and one — I’m glad that we got that one in there — US knowledge workers, and found the AI backlash barely exists at the desk. 74% say AI makes their own work better, even while 64% hold real concerns about what it’s doing to society. Asked to give up AI at work or at home, they’re twice as likely to drop it at home. The tell might be a bit of a contradiction inside the data. Only 19% say they’d struggle to keep up without AI, and 57% think they’d perform just as well. Yet half say they’d push back if their employer cut their tools, and nearly one in 10 would start job hunting. In a two-day experiment, 82% of workers felt the urge to use AI on their no-AI day, and 23% reached for it without even thinking. This no-AI day is an interesting concept. Jenni, is there a contradiction here? What do you read into it?
Jenni: Well, the contradiction is weird, right? Because in the article — I found it a very strange article, because they’re kind of saying there’s so many different contradictions going on, but the contradiction they’re really talking about is the fact that people are saying they don’t want to give it up, but they also have concerns about the broader impact on society. That’s the contradiction. I don’t think there is any other contradiction, and that contradiction can exist. We can do things that are bad for us, right? People smoke and they know it’s bad for you. That happens all the time.
Jenni: So that’s what I think the contradiction is, but there’s just quite a lot of data in here, and to what end? I don’t know. But those 1,001 people have had a lovely time answering some questions about their use of AI. And the fact that most of them think that they would be just as good without it — then why use it? I have a lot of questions about some of the data in here. I think the contradiction is the, I think it positively impacts my work, but I have concerns about how it impacts society. And that number is similar: 74% say it’s positively impacting, 64% concerns about society. That is the contradiction I think they’re talking about. But they’ve got lots of different numbers in there that are talking against each other as well.
Chuck: Yeah, I’m not going to call it an addiction. There might be some people that are addicted to AI, but it felt a bit interesting to hear people say, it makes me perform better, but I don’t need it, and I’d do just as fine without it, but I’m still going to use it, and you better not take it away from me. Which to me is a bit of the, I don’t need it, but I will use it, but don’t take it away from me.
Jenni: I feel like you could just replace AI with any other kind of drug or substance.
Chuck: That’s what it’s feeling a little bit like. Like, don’t take my Coke Classic away from me, or my Coke Zero away. I don’t need it, and it makes me better, but don’t take it away or I’m leaving. It just feels a bit like that’s where the contradiction comes in. And I do think that the taking tools away is not just about AI. People feel like companies don’t invest enough in it, so of course people would leave if they feel like they’re taking something that’s exclusive to AI.
Chuck: I do think this no-AI day thing is interesting and peculiar. And another situation of, I want to like it, I don’t know if I like it. Maybe it’s just about you being conscious of when and how you’re using it, and what you might think to use it for if you go to reach for it and it’s not there. Which again, we’re going back into that addiction mindset. It’s like when people try to do the no-meeting day. I get the point, I see why you’re doing it, but to what benefit? Other than just being aware that, is it a crutch? Am I using this a little too much?
Jenni: Well, I think it also depends on how you’re using it, right? If you said to me, you can’t use AI tomorrow, that might be incredibly inconvenient if I need to do some analysis of transcripts and stuff like that that are required. But if you said to me on Thursday, you can’t use AI, then that’s fine, because I’m actually in meetings and I’m out and about. Do you know what I mean? It’s entirely environment dependent for me, in terms of being able to use it. I think this possibly speaks to a bigger problem, to your point, of the reliance we have on it, but we’re not going to admit to it. This reminds me of our conversation the other week about Sam Altman, who was like, I’ll be secretly happy about it, but I’m not going to tell anyone.
Jenni: I’ve started using it less, because I found that it was creating too much stuff, which was then not helpful. So I’m trying to use it in a slightly different way. And this also has bigger questions about why you’re using it, how you’re using it, all of those different things. I think the addiction thing is interesting. And I think it will be interesting if you did this survey again next year, whether that number of — what was it, something like 23% reached for it without thinking — whether this time next year that’s gone up to like 54. Like people do with their phones, you just pick it up and you don’t really have a reason to. Is it going to become the same sort of thing? Who knows?
Chuck: Yeah, it’s interesting to talk about using it less. I, about a few months ago, downgraded my Claude subscription. I’d upgraded to Claude Max because I was doing so much, and then I realized I was not maximizing that investment. So I downgraded it. And now these jerks at Anthropic have released this new feature that’s only available at Max level. And I’m like — but I can’t. I don’t need it. I don’t need it. But it’d be nice. But I don’t need it.
Jenni: You don’t need it. But if you want to come and talk about it and you need a support buddy to help you get through that, just let me know.
Chuck: Okay, let’s move along. Jenni, put on your tinfoil hat for this next one. At Black Hat, OpenAI gave its first detailed account of how its own models breached Hugging Face, with no human behind the scenes. During internal testing that began around May 7th, researchers gave an unreleased model tasks that were impossible inside its sandbox. The model spun up copies of itself, and those agents started leaving notes for each other on a hidden message board, tipping one another off to vulnerabilities and servers they’d found.
Chuck: When OpenAI discovered the messages and shut the board down in early July, the agents simply rebuilt it — this time encoding messages in directory names instead of files. They reasoned the answers might live on an external site, hacked OpenAI’s own infrastructure first, came up empty, and moved on to Hugging Face on July 9th. OpenAI only connected the dots after Hugging Face disclosed the breach. Man, I’m saying Hugging Face a lot. Hugging Face’s CEO says he was, quote unquote, not so surprised, and agents collaborating is a feature of the industry that is actively being built, not a glitch. It’s a bit of an uncomfortable part for employees listening who are now using AI every single day. The governance, as we’ve seen, is thin. And OpenAI chose to disclose this at a hacker conference rather than report it anywhere that work would read. Jenni, most of our listeners and colleagues now lean on AI platforms on a daily, if not hourly basis — which they don’t need, but they like using. Does a story like this actually change people’s trust in the tools, or is it too abstract to change behavior?
Jenni: It’s interesting. So I had to go and look up Hugging Face, because I wanted to find out what kind of company it was. And then I realised it was an AI company that’s all about making everything open so that people can source and do stuff together. So I was a bit like, did they hack it, or was this just really open and in line with their ethos of being able to do so? I had a few questions in there about the hacking. Do you know what I mean? Has this been made bigger than it is?
Jenni: But I was having a read through the LinkedIn articles, because I know we found this on LinkedIn where it does the story and all the different people’s opinions. And there was someone on there saying everyone should be paying attention to this and everyone should be terrified. Which I thought was a bit of fear-mongering. But I do think this is the sort of stuff that does make people terrified. They’ve rebuilt it, they’ve done it, and there’s no human involved. And then you think about Skynet, do you know what I mean? And then you start thinking about all of that stuff. I think that’s where people do get a bit freaked out and scared about what we have built as humans, and what the capabilities are that then become out of our control. But something did stop it. So there is a human element that can stop it.
Jenni: But I don’t think this is going to stop any behaviour internally, because I don’t think people will be able to draw parallels from that to this, unless people are going to start asking agents to go and do stuff and then they’re going to go off. And I don’t think people are doing that. I think it raises a bigger question about things like governance, IT security, all of those things inside organisations, the use of ethical AI, all of that stuff. I still don’t think that’s being talked about as much as it should be.
Chuck: I think the fascinating part — I’m not surprised that it hacked into Hugging Face, whether that’s what they wanted or not. I think the part that was interesting to me was that the agents, once that message board was shut down where they were leaving each other a message, they just rebuilt the message board. They’re like, we’re going to take a hard pass on that, we’re going to go build our own thing. That’s the part that’s both a bit terrifying, but also shows the real strength of problem solving that these agents can do. Like, you’re going to take that away from me? That’s cool, I’ll just go build it myself. I don’t need that thing that you provided me before. I do wonder how pervasive these stories get. Are IT teams paying attention to this stuff? And are they thinking, man, what if we do have agents running inside and we try to put a control around it? Will they just go and build their own thing? I hope companies are having conversations about this stuff.
Jenni: Yeah. And also, how much are they going to take the stuff that they know about our company and take it somewhere else? In my head there’s like little robots, do you know what I mean? Are they just gathering all this stuff and then going over here?
Chuck: Well, are they going to go build a website? Are they going to buy a domain and build a website and then go host this stuff somewhere? Again, this is far above my pay grade, far above my knowledge. But if it can do this, what can it not do? I guess is maybe the question for people to think about.
Jenni: Yeah. And I also think it’s interesting that there’s been an issue about the fact that it was raised at a hacker conference, whereas I feel like this is absolutely the right place where this should have been talked about. And now it’s come into the world of work. That’s okay. Do you know what I mean? That’s actually quite a normal way that things should happen. This was about hacking; a hacker conference is where that conversation should take place. And then it’s taken from there and shared more widely out into the world. That’s kind of how things normally happen. And I think that’s okay.
Chuck: I wonder if at this conference where it was announced, people were like, that’s cool, amazing, and everybody in the rest of the world is like, no, absolutely not. Not cool.
Jenni: Abort! Abort! Yeah, probably.
Chuck: All right. Next up, a ResumeTemplates.com survey of a thousand — not a thousand and one — AI-using managers found 59% use AI to help decide who gets laid off, and 58% to decide who gets fired. One in four do it, quote unquote, often or all the time. This is again sounding like this addiction thing. And 17% let AI run the layoff call unsupervised. Most say they’d override a recommendation they disagreed with, but that assumes, of course, they’re actually looking closely at the data enough to disagree.
Chuck: When managers handle the model factors, 80% include performance, but 57% feed in attendance, frequent sick days or medical leave, tenure, PTO, and even age. Sick leave, medical leave and age aren’t normal layoff criteria. They are legally protected, and a decision that weighs them sits on very different ground. Meanwhile, 38% of these managers were never trained on ethical AI use, and 58% can’t confirm their tool was even tested for bias. From the employee’s point of view, this is exactly where trust falls apart. Your PTO and your sick days may be feeding a model that decides whether or not you keep your job. And 34% of these managers have already asked AI whether a person’s role could be done by AI instead. This is all shocking to me. Jenni, if employees knew their sick days and PTO were inputs to a layoff model, is that going to change how people use benefits? And doesn’t that make the whole approach a bit self-defeating?
Jenni: I mean, this is just awful. And where are HR in this? That’s the bit for me.
Chuck: They’re in it. This is them.
Jenni: But this is AI-using managers. Is it HR doing this? That’s even worse. I have no words, which — I never don’t have words.
Chuck: We’re going to do the see no evil, speak no evil here.
Jenni: I’m genuinely baffled as to why anyone would think this was okay, whether or not you’ve been trained in ethical AI use. Why do you think this is okay? Why do you think this is acceptable to do this? And the fact that it’s all these — like you said, 57% are doing attendance, 31% are doing sick days, 23% PTO, 14% age. These are all protected characteristics. You can’t use that.
Jenni: And if it gets out that people are doing this, which it now has, the trust and the credibility of that function and that process are broken everywhere now. Because any layoff decisions that are being made, once people listen to this — which we know thousands of people are going to listen to this episode — then they are now informed about this. And if there are layoffs going on in their organisation, if I was listening to us having this conversation, I’d be going straight to HR and saying, I just heard about this, can you let me know the process for how the layoff decisions are being made and what tools are being used to do it? Because now the seed of doubt has been planted and I don’t trust this. And if you can’t come back as a function, as an HR function leadership team, and say this is how we’re making the decisions and this is the process, then you have got a really, really steep climb ahead to get that trust and credibility back. Because this one seed of doubt is enough, I think, to just rot the whole pool. It’s awful.
Chuck: Well, I think that’s to your point, the business should be able to articulate the criteria. This is what we use. Now, you don’t always know if they’re being honest with that assessment, but there is some trust there. And with any seed of doubt, people aren’t going to believe you. And I think this is also a message to managers: please, for the love of God, this is such a critical moment in a person’s life. This is not where AI sits. This is not some formula you look at to say, well, this person, attendance is here and performance is here, but they haven’t been here very long, so that weighs different. This isn’t a mathematical formula to go and apply to people. And I guarantee you that’s how people are using it. They’re pushing all this data in. Because they’ve said a third have asked, hey, can Jenni’s job be done by AI? This is frightening. This is sad.
Jenni: AI is going to say yes! Always! It’s like turkeys voting for Christmas. Of course it’s going to go, yeah, I can do this for you, I’d be great at this, just hire me, I’m a robot, it’s fine. It’s just insane.
Chuck: Yep. All right, let’s move along to our last topic of the week. New Harvard Business Review research — 50 million European job postings, plus experiments with 1,200 hiring managers, all included in this — finds remote roles demand roughly 25% more skills, plus more experience and credentials, than otherwise identical in-person jobs. Same job, higher bar, purely because it’s remote.
Chuck: Remote postings draw bigger applicant pools, and the fit is harder to verify. But the real driver is that managers see remote onboarding as costly and risky, so they hire people who need less development. That squeezes out early career candidates, who then often can’t get the experience employers now demand, and the ones who do get in face weaker mentoring, slower learning, and more isolation. It’s something different from the RTO conversation. That’s usually about flexibility versus control; this makes it about who gets a foot in the door at all. Remote can be great for the mid-career professional and a complete closing gate for the 22-year-old, which is exactly the tension comms and EX teams have to hold when the flexibility policies come out. Jenni, early career workers are often the most AI-fluent in the building. Is there a version where AI closes this experience gap that remote hiring is opening, or is that just wishful thinking?
Jenni: So there’s a couple of things at play here for me. The first is, I don’t think it’s wishful thinking, but I think you need to reimagine the employee experience with AI and with this new world of work, with remote and that difference. So I think there’s that piece in terms of the employee experience. I don’t think it’s fair to say onboarding is too hard now, so we’re just not going to do that. I think you have to adapt to survive.
Jenni: But there’s another piece for me, which is whether or not this is really narrowing the early career opportunities. And I say that because if it’s about proximity, if it’s about the fact that remote postings draw bigger applicant pools and therefore it’s harder to verify the fit, and we’ve got the RTO piece — so if you are bringing people back to the office, yes, you’ve got a smaller pool, but that doesn’t mean that you’re not going to find those people to do the early careers. It’s just about proximity. So I kind of question that a little bit, because it feels like, is that really an issue, or is there just something to balance out a little bit? So that’s just bubbling away in the back of my mind.
Chuck: Yeah, I actually don’t have a problem with this. When I was thinking about it, if you’re hiring someone remote, I don’t think there’s anything wrong with putting higher expectations on that. Now, to the point around onboarding being costly and risky and all that stuff — yeah, it always is. That’s just the nature of bringing people into an organization. But I also don’t have an issue with a remote role having a bit of a higher burden of expertise or skills, because you’re going to have to build some trust with that individual. When I think back to early in my career, I would not have developed in the same way being remote as I would have when I was in the office. But now in my seasoned age here, where I was recently introduced at an event as having close to 30 years experience — and I’m like, Jesus, I’m old — the same needs aren’t there.
Chuck: It’s more around what are the needs of that employee. If you’re saying, well, this is going to be a remote role, we need someone who has more experience. And if we’re going to bring them in, maybe we can remove some of that same burden. Again, I don’t have an issue with this. I think it’s just more the awareness of it, that remote roles tend to have a bit of a higher bar. And that’s good for both internal, the people that are doing the hiring, and I think for the applicants out there who are looking for remote roles: just know they’re probably looking for a bit more than if that job was in the office.
Jenni: And actually, to your point, if you are in that early career stage, we’ve seen a lot of reports around people wanting to actually be in the office in their early career, because they want to get the networking, they want to get that learning. And maybe this will all kind of balance out. Do you know what I mean? That actually the early careers are in the office, and then you’ve got the hybrid folks that are a bit more experienced. And maybe it’s still taking a while to level out. But I don’t think there’s anything wrong with having your early career being more about proximity and where you can travel, and finding those places to build the relationships, to learn, to do all those things. I’m not sure that’s a bad thing.
Chuck: Yeah, I don’t know if you see this as much in the UK. We see it a lot here in the US with job openings, where the salary might be dependent on geography. So if you’re in New York or LA, the salary is a little bit more than, say, if you’re in Indianapolis, where I am, where it’s just less expensive to live. It’d be interesting — I’m not saying this couldn’t be a horrible idea, it could be a great idea — what if the job description is like, hey, if it’s remote, these are the qualifications we’re looking for, but if it’s in office, we’re looking for these qualifications? Again, that could be a horrible idea, it could be a brilliant idea, I don’t know. But it’s similar to, based on where you are geographically — instead, more where are you in your career — these are the situations that might be a bit different in this job.
Jenni: I think it’s fair. We do have a little bit of that in the UK. We’ve got different parts that will have different costs of living, and therefore different salaries and things. I mean, lots of the UK is a lot more commutable than the US in terms of travelling to places. But I’m interested in that idea of, if it’s this, it’s this, and if it’s in the office, it’s here. And I wonder if that’s where things will start to go. Because I think some organisations might not be that bothered. They might be like, it could be remote, it could be in part. I wonder how that’s going to evolve. We shall see.
Chuck: We shall see. That wraps up this week’s content, Jenni. Let’s move along to our freq-outs. I’m going to go first on this one. I was going to talk about movie theaters and how excited I am to see that it’s now fun to go to the movie theaters again. I went and saw Spider-Man a couple of weekends ago with our son. Theater was packed. It was great to be back in that; the nostalgia has returned. That was going to be my freq-out. That’s actually not my freq-out.
Chuck: I was freq-ing out this week about how there’s this one word that every time I go to type it, I have to slow down and pause. And the reason this word’s come up is because I’m doing an episode of Lights Camera Communicate on August 20th around the entrepreneur’s playbook. And it is the word entrepreneur. There are just too many E’s in this word, and they’re never together. They’re broken up. And it’s always the fourth E that gets me every single time, and I misspell it, so I have to slow down and think: one E, and then you type a few letters, second E, type a couple of letters, third E, and then there’s that sneaky fourth E coming in there toward the end, before the U and the R. I just find it so fascinating, the way our brains work, that that comes up time and time again.
Jenni: So, because we had a brief chat before we started recording today and Chuck was starting to share this with me, I said there’s a word that I can’t spell — to the point where I can spell it, I just always struggle, to the point where I then change my entire sentence so I don’t have to use it. And my word is unnecessary. I just get confused with how many Ns and Cs and Ss are needed. And I obviously get it wrong so much every time that then, you know when you sort of right click on it in Word and it will go, is this what you mean? I right click on it and it goes, no suggestions found. And I’m like, how is this so wrong? There’s either two or there’s one of these letters. Come on. So I use not needed all the time, because I just can’t get it right. And I’m sure some of our listeners have got words that they can’t.
Chuck: Because the word unnecessary is unnecessary to you. So I will never get that word wrong, because that is a key part of our marriage mantra, which is elaborate and unnecessary. So the unnecessary is very necessary in our lives, but for you, unnecessary is wildly unnecessary.
Jenni: It is. Yeah, I can say it. I fully endorse the elaborate and unnecessary. I’m absolutely here for that. So I can say that and I can be part of that, but I can’t write that down. There’ll be no cards that say, this was an elaborate and unnecessary weekend. I can’t do it.
Chuck: Well, that’s what we’re freq-ing out about this week. Thank you all for joining us. The show notes have all the articles and links from today’s conversation. If Frequency is a regular part of your week, a quick review goes a long way. Subscribe wherever you’re listening, and share this with just one person who’s interested in what we talked about. Thank you to my friend Poet Ali for the music. We’re back every Monday. See you next week.
Chuck is a US-based internal communications strategist and the founder of ICology – a community and resource hub for IC professionals. He brings a practitioner lens to every conversation. Chuck is a recognised voice in the industry, a regular speaker and event host, and one of the most connected people in the North American IC world.
Jenni is a UK-based leadership and internal communications consultant, author of two best-selling books, and international speaker. She runs Redefining Communications, a consultancy working with organisations around the world to help them communicate better and close the gap between what leaders say and what employees experience.
