In partnership with
📋 5-3-1 — 5 Culture Plays on AI and Job Security, 3 Playbooks, 1 In The News
Gallup's latest poll finds that 27% of US workers worry that technology could make their job obsolete. That's a new high, up from 20% last year and 13% in 2017. Last week a Washington State University study put a cost on the worry: workers who feared most for their occupation's future were more likely to justify bad behaviour at work, and to report slacking off, turning up late and taking a colleague's lunch.
What caught our attention is what's missing. Gallup says the worry is running ahead of actual job losses from AI, so the fear is there whether or not anyone's role has changed yet. Mercer's 2026 research found that only 19% of HR leaders say they consider the emotional and psychological impact of AI when they plan their rollouts, while 62% of employees feel leaders underestimate it. We went looking for an employer that publishes how worried its own people are about being replaced. We couldn't find one.
So this week we dropped the guest and went after that gap. Five real cases, from banks, a state government and a healthcare group, on what happens when AI and jobs are announced badly or written down carefully. Three Playbooks on measuring the fear and answering it. And the study itself, with the numbers around it. Two of the five cases are cautionary tales, and where a source shows what someone intended and not what happened next, we say so.
Welcome to the eighteenth issue of The Work Life Reporter. This week you get:
5x Culture Plays on AI and job security, from banks, a state government and a healthcare group
3x Micro-Playbooks for measuring the fear and answering it
1x In The News: the Washington State University study and the numbers around it
1x Reader poll: last week's update and one new question
Let's get into it.
Speak naturally. Send without fixing.
Wispr Flow turns your voice into clean, professional text you can send the moment you stop talking. No rough transcription you have to clean up. Actual polished text, ready for email, Slack, or any app.
Speak the way you think. Go on tangents. Change your mind mid-sentence. Flow strips the filler, fixes the grammar, and gives you text that reads like you spent five minutes writing it.
THE FIVE CULTURE PLAYS
Play 01
Standard Chartered Put a Number on AI and Jobs in a Press Release. Then Its CEO Had to Apologise for How He Described the People
On 19 May 2026, at an investor event in Hong Kong, Standard Chartered published its targets for 2028. One of them was to raise income per employee by about 20%, "aided by a reduction in corporate functions roles of >15 per cent by 2030". Press reports put that at more than 7,000 roles. Asked about it in the room, CEO Bill Winters said it wasn't cost cutting. It was "replacing, in some cases, lower-value human capital" with the money the bank was putting in, according to Banking Dive, and his choice of words spurred what the publication called "blistering criticism".
By the Wednesday Winters had sent staff a memo saying that where roles fall away, it reflects "changes in the work, not the value of our people". On the Friday he apologised on LinkedIn, saying his remarks "caused upset to some colleagues". He also said the bank has invested for years in helping colleagues whose roles may be displaced to build new skills. The number was in a press release and caused little fuss. The row was about one sentence said in a room.
People hear the sentence about them, not the table around it. If a plan puts a number on roles, the words used about the people in those roles are part of the plan.
TEMPLATE: THE BEFORE YOU ANNOUNCE CHECK
| The Before You Announce Check 1. Write the number and the timeline in one plain sentence. Then read it as someone in an affected role would. Does it sound like a fact about the work or a verdict on the person? 2. Cut any word that ranks people by value. "Lower value", "surplus" and "low performers" will be quoted back at you. 3. Say what you'll do for the people in those roles on the same day: retraining, redeployment, notice and who to ask. 4. Brief managers before staff read it in the press, with a short list of what is decided and what isn't. 5. Plan the second message. Decide what you'll tell staff on day two and who answers their questions. |
Play 02
Commonwealth Bank Said AI Made 45 Jobs Redundant. After a Union Challenged It, the Bank Said It Had Got That Wrong
In July 2025 Commonwealth Bank of Australia told staff that 45 customer service roles in its Customer Service Direct business were no longer needed, because an AI voice bot had cut call volumes. The Finance Sector Union said the opposite was happening: calls were going up, overtime was being offered and team leaders were taking calls. It also said the bank kept refusing to be open about the figures, and it escalated the dispute to the Fair Work Commission.
On 21 August 2025 the bank reversed the decision. A spokesperson told Information Age that its first assessment "did not adequately consider all relevant business considerations" and that "this error meant the roles were not redundant". The bank apologised to the employees concerned. According to the union, those affected could stay in their role, move elsewhere in the bank or take the redundancy. The call volume claims are the union's, and the bank's own statement says only that its assessment was incomplete. Even so, a cut blamed on AI was reversed within weeks, after the union pushed for the numbers.
If AI is the reason for a cut, you should be able to show the work that proves it. If you can't, the people doing that work will check for you.
TEMPLATE: THE SHOW YOUR WORKING CHECK
| The Show Your Working Check 1. Write down the measure that shows AI has reduced the work, such as calls per week or tickets per hour, and the dates it covers. 2. Share it with the affected staff and their representatives before the decision is final. 3. Ask the people doing the work what they see. If their account differs from the data, find out why before you announce anything. 4. Set a date to look at the numbers again once the tool has been running for a few weeks. 5. Decide in advance what you'll offer if the check shows you got it wrong: stay, move or leave, as the bank did. |
Play 03
Bank of Ireland and Its Staff Union Wrote Down an AI Agreement, and the Union Has Measured How Worried Its Members Are
In March 2025 Bank of Ireland Group and the Financial Services Union published a short AI agreement covering colleagues in Bands 1 to 3 who are covered by collective bargaining. It says that with AI, "humans will be kept at the centre of work organisation". It describes joint planning on retraining as "a long term job security measure", says employers should provide the training needed for new technology during working hours, and reaffirms the bank's existing job security and change management agreements. Either side can end it with six months' notice.
It runs both ways. The agreement says employees need to be flexible about redeployment, and someone who refuses a reasonable alternative role is handled under the existing change management agreement. It isn't a promise of no job losses, and it contains no numbers on training or redeployment. What it has is a process. The union also did something we couldn't find an employer doing: it measured the worry. A survey of 604 of its members found that 88% believe AI will lead to job displacement and 60% feel less secure in their roles than they did five years ago. Workers under 35 and those in administrative roles were the most anxious. It's the union's own survey of its own members, so read it as a signal and not a national rate.
A promise with a process attached is harder to argue with than a promise alone. Say who is involved, when training happens and how either side can step back.
TEMPLATE: THE ONE PAGE AI AGREEMENT
| The One Page AI Agreement 1. Write the principle in plain words, for example that people stay in control of how work is organised. 2. Name the job security and change agreements you already have, and say the AI agreement sits inside them. 3. Put retraining in working time and call it part of job security. 4. Say what you expect from staff too, such as flexibility on redeployment, and what happens if someone declines a reasonable alternative role. 5. Give both sides a way out: a notice period and a dispute process. |
Play 04
Pennsylvania Put in Writing That It Won't Use AI to Substitute for State Employees. Its Union Leader Says That Isn't a Guarantee Forever
In March 2025 Governor Josh Shapiro published a letter to state employees setting out eight principles for using generative AI at work. The first says AI will be adopted to help employees be more efficient and productive, "not with the agenda of substituting Commonwealth employees with AI". The second brings union covered employees into future pilots. The third says that if AI saves time, the state will look for ways for employees to reinvest it in better service. SEIU Local 668, which represents 10,000 state workers, reached a separate agreement with the state that includes a human in the loop approach and affirms that a public employee is a person.
By August 2026 more than 3,000 state employees were using AI tools, up from the 175 who piloted them, with another 6,000 in training, according to the Philadelphia Inquirer. The state says AI has improved efficiency without eliminating jobs. Pilot users said they saved an average of 95 minutes a day, a figure the staff themselves reported. The most useful line comes from the union. SEIU Local 668 president Steve Catanese told the Inquirer that most members have "some serious concerns about what it means for their future in the long term". Asked about the protections the union won, he said, "Is it fully guaranteed forever? No." He added that they are a level of protection almost no other public servants in the country have.
Be specific about what you're promising and honest about what you aren't. The most believable line in this story is the union leader saying it isn't forever.
TEMPLATE: THE PROMISE WITH ITS LIMITS
| The Promise With Its Limits 1. Write the promise as one sentence about what AI is for and what it isn't for. 2. Name who sees pilots before rollouts, including staff representatives. 3. Say where saved time goes, so people don't have to guess whether it means more work or fewer colleagues. 4. Say what the promise doesn't cover. Anything you leave out, people will fill in themselves. 5. Set a date to review it and report back. |
Play 05
Kaiser Permanente and a Union Alliance Wrote the Employee Voice Into Their AI Agreement, With a Task Force to Hear It
In its 2025 national bargaining with the Alliance of Health Care Unions, Kaiser Permanente agreed text on AI and technology. It says "the employee voice is essential to the successful adoption of new technology", and that employee input will be built into the process, from spotting the problem through to training and evaluation. Both sides also commit to "timely communication to the workforce about new technology that impacts employees' roles and workflows".
To make that real, the text sets up a joint task force on technology and AI within 120 days of ratification. Its job is to share planned rollouts, learn together and provide "a forum for early engagement". Topics can include hiring, performance evaluation and corrective action. Kaiser also agrees to tell the union, following local practice, about any new AI used to monitor productivity or performance. The AI section doesn't add a promise of no layoffs. It points to agreements that already exist on union security, employment income security and workforce planning. The text we read is a draft published by a union, so the final wording may differ, and there are no results to report yet.
A promise to listen is easy to make. A standing group with seats on both sides and a start date is something you can check.
TEMPLATE: THE STANDING SEAT CHECK
| The Standing Seat Check 1. Name the group, who sits on it from each side and the date it first meets. 2. Agree what it sees before rollout: planned tools, what they're for and who they affect. 3. Say what it doesn't replace. Here, local bargaining still applies. 4. Commit to telling staff about any tool used to monitor productivity or performance. 5. Report back at set dates on what changed because the group met. |
THREE MICRO-PLAYBOOKS
Playbook 01
Measure the Worry, Not Just the Usage
HSBC's CEO Georges Elhedery was clear about what he'd be watching. In a company transcript dated 2 April 2026, talking about the generative AI tools going out to staff, he said "I will be monitoring usage" and that "utilisation is not optional". Usage is easy to count. The worry is harder, and it's where the cost sits. Gallup compared each US worker with themselves over time, between 2023 and early 2026. When someone's fear of AI displacement rose, their engagement and job satisfaction fell, their burnout rose, and they were 4.1 percentage points more likely to be job hunting.
Gallup's own advice is to "monitor engagement as a leading indicator", because a drop may be an early sign of AI anxiety even when people don't call it that. You can also just ask. We'd add one question to your next pulse survey in the shape of Gallup's: how likely is it that your job will be eliminated in the next five years because of new technology, automation or AI? Then add a second: do you know what AI will and won't change about your role? Report totals only, and don't show any group smaller than ten people.
Mercer's 2026 survey found that only 19% of HR leaders consider AI's emotional and psychological impact in their digital implementation strategy. Asking is how you start to.
Playbook 02
Have the Conversation Before the Rumour Does
Gallup's September analysis says to treat the AI rollout as "a communication challenge as much as a technical one". Displacement fear isn't driven by how exposed a job is alone. It also comes from whether people get clear, credible information about what AI means for their role. Workers who strongly agreed that their organisation cares about their wellbeing, or that they feel respected, were about 6 to 7 points less likely to report displacement concern. Gallup's suggestion is direct conversations between managers and their people.
In the UK, Acas and YouGov asked 1,023 employees in March 2025 about their biggest concern with AI at work. Job losses came top at 26%. Acas told employers to have clear AI policies, to consult employees and their representatives on introduction, to explain how AI can improve roles and to reassure staff that human involvement will still be needed. JFF's survey, fielded in late 2025 with 3,020 respondents, found that 56% of workers say their employer hasn't consulted them about how AI tools are used in their work.
Three questions work well in a one to one. What have you noticed AI doing to your work so far? What would you like to know that you haven't been told? What do you need from me? Then make one promise you can keep: say what you know, what you don't know yet and when you'll next update them. Don't promise anything about the future you can't deliver.
Playbook 03
Look Where the Fear Is Highest First
The fear isn't spread evenly. In Gallup's August poll, 34% of workers aged 18 to 44 worried technology could make their job obsolete, against 19% of workers aged 45 and over. JFF found that 40% of early career workers had changed or were considering changing their career plans because of AI, against 19% of more experienced workers. The Bank of Ireland union's survey found the same pattern among under 35s and administrative staff.
Exposure matters too. In the Washington State study the link between worry and bad behaviour was stronger in heavily exposed industries such as finance and insurance. Lead author Tahira Probst's explanation is about alternatives: a bartender whose bar closes can find another bar, but a medical transcriptionist whose work has been automated can't simply do the same job somewhere else. And Gallup found something that cuts against the idea that using AI more settles people down. In most survey waves, workers who use AI daily or several times a week were more than twice as likely as less frequent users to say their job is very likely to be eliminated within five years, though only about 3% to 4% of workers said so overall.
So cut your own results by who is most exposed, not just by department. List the tasks in each role family that AI tools can already do, start with the highest, and pair early career staff with one named person who can answer their questions. Gallup found that strong organisational support matters most for workers with the fewest ways to reassure themselves: those with less seniority, fewer outside credentials and shorter track records.
IN THE NEWS
NO GUEST THIS WEEK, ONE STORY INSTEAD
A new study links the fear of being replaced by AI to workers cutting corners, and Gallup says that fear is at a record high.
Researchers at Washington State University published a study in Occupational Health Science last week, led by psychology professor Tahira Probst with graduate student Rebecca Lindgren and researchers in Italy and Belgium. It looked at workers across 18 industries, at how worried they were about their occupation's future and at how they behaved at work, and compared that with how exposed each industry is to AI. Workers who worried more were more likely to morally disengage, which Probst describes as "basically a way of justifying bad behavior". They also reported more counterproductive behaviour at work, such as slacking off, browsing the web on shift, arriving late and taking a colleague's food. The pattern was stronger in heavily exposed industries like finance and insurance and weaker in ones like forestry and construction.
Probst's explanation is about whether there's somewhere else to go. A bartender whose bar goes under can find another bar, she says. For a medical transcriptionist whose work can now be automated, that isn't the case. She points out that this kind of behaviour hurts colleagues and wastes organisational resources, and the university's release says employers should pay attention to how they discuss and roll out AI.
Gallup supplies the bigger picture. Its August 2026 phone poll of 1,200 US adults, 507 of them working, found that 27% of workers worry technology could soon make their job obsolete. That's a new high. In Gallup's words, "workers' growing concern appears to be outpacing actual job displacement due to AI". A separate Gallup analysis of its panel data, which tracks the same workers over time, links rising displacement fear to lower engagement and job satisfaction, higher burnout and a greater chance of job hunting.
That's the part we keep coming back to. The fear is real, it has a cost you can measure, and it's moving faster than the thing it's afraid of. And yet Mercer found that only 19% of HR leaders say they consider AI's emotional and psychological impact in their implementation plans, and JFF found that 56% of workers say their employer hasn't consulted them about how AI tools are used in their work. We went looking for an employer that publishes how worried its own staff are about being replaced. We couldn't find one. The closest we found were pollsters, a union survey of its own members and a CEO who said he'd be monitoring AI usage. That doesn't mean no employer is measuring it. It does mean the number isn't being shared, which is worth asking about.
How to read this. We read the Washington State University press release and coverage of the study, not the full paper, so we can't give you its sample size. It reports a link between worry and the behaviour people reported about themselves, in industries ranked by how exposed they are to AI. It doesn't show that AI caused anyone to behave badly, and it measures potential exposure, not whether employers have actually adopted AI.
FIVE NUMBERS WORTH KNOWING
27% of US workers worry technology could make their job obsolete
Gallup, August 2026 poll of 1,200 US adults, published 15 September. A new high, up from 20% in 2025 and 13% in 2017. Among workers aged 18 to 44 it's 34%, against 19% for those aged 45 and over.
40% of employees are concerned about job loss due to AI, up from 28% in 2024
Mercer's Global Talent Trends 2026, a survey of nearly 12,000 executives, HR leaders, investors and employees worldwide, published in January.
Only 19% of HR leaders consider AI's emotional and psychological impact in their implementation strategy
Also from Mercer, which found that 62% of employees feel leaders underestimate that impact. Mercer's word is "consider", not "measure", so this is the closest number we found to a measurement gap and not proof of one.
56% of workers say their employer hasn't consulted them about how AI tools are used in their work
JFF survey of 3,020 respondents aged 16 and over, fielded 28 November to 8 December 2025 and published in March 2026. Only 39% of workers said they feel optimistic about AI's impact on workers, down about 10 points on 2024.
41% of UK employees fear AI's impact on their jobs, against 25% of HR decision makers
YouGov for The Access Group, 1,000 UK employees and 503 HR decision makers, fieldwork 11 to 19 December 2025, published in February 2026. Each group was asked about their own job, so this isn't what HR thinks employees feel, but HR leaders were less worried than the people they support.
Sources: Occupational Health Science and Washington State University, Gallup, Mercer, JFF and The Access Group. Reply if you'd like the links.
READER POLL
Last Week's Poll, and This Week's Question
Last week we asked which is harder for you right now: knowing what your people really think, or doing something about what you already know. We promised to share the result, and we'd rather do that properly than hand you a number before it means anything. So we're keeping that poll open for one more week and we'll share both results together next Monday. If you missed it, you can still answer it in Issue 017. Totals only, no names and no companies.
This week's question comes straight from the story above. One click is all it takes, right here. We're also running the same question as a poll on our LinkedIn page, so you can answer there if you'd rather.
Does your organisation measure how worried your people are about AI replacing their jobs?
Next Monday we'll share the results of both questions, with the newsletter answers and the LinkedIn answers shown separately and together, including how many of you answered and no names. You can read how we handle your answers on our privacy page. theworklifereporter.com/privacy
THE WORK LIFE REPORTER LIVE
A weekly LinkedIn live series where we take the most interesting conversation from the newsletter into a real room, with guests, debate, and the questions the newsletter doesn't have space to answer.
This Week
No Episode This Week. We Spent the Slot on One Story Instead of a Guest
Back to a normal live conversation soon.
If there's an episode you missed the first time, this is a good week to catch up on it.
And that's a wrap for Issue 018. No guest this week, one story instead. Hit reply and tell us whether your organisation has ever asked your people how they feel about AI and their jobs. We read every answer.
Follow The Work Life Reporter on LinkedIn for what we share between issues. TWL Live returns with a guest soon. New here? Subscribe free at theworklifereporter.com.
PS. Know someone who leads people and would find this useful? Forward it on. Ten seconds to share, a lot longer to write.
PPS. Haven't answered the poll yet? It's one click, up above. We'll share the result next Monday.