News & Insight

About 35% of current jobs in the UK are at high risk of computerisation over the following 20 years, according to a study by researchers at Oxford University and Deloitte. Go to http://www.bbc.co.uk/news/technology-34066941 and type your job title into the search box below to find out the likelihood that it could be automated within the …

Continue reading “Will a robot take my job?”

When Jensen Huang told Nvidia employees on November 8 that he had been “unfaithful” to his wife, there’s a phrase that quickly became omnipresent in the media: Huang had been the victim of a “honey trap”, laid by Alice Huang (no relation), a Singapore-based tech executive identified as a Chinese spy by U.S. intelligence agencies.

“Honey trap,” seriously?

The story behind the phrase is that a good, hardworking man was minding his own business when a serpent of a woman lured him off the righteous path by dishonest means and propositions he, being a man, simply couldn’t refuse. It is assumed that male sex drive is some kind of force of nature that no adult can reasonably be expected to control. So there’s a double sexist stereotype here. Can a 63-year-old billionaire CEO really be supposed to have no say over his sex life? Like a bear wouldn’t be able to resist honey?

Anyone who, like me, grew up in the 1980s knows a close relative of this idea: the “homewrecker,” a word aimed at the other woman and never at the husband who cheated. One of the decade’s most successful films, Fatal Attraction (1987), tells the story of a married man (Michael Douglas) whose affair turns into a nightmare when the woman (Glenn Close) refuses to stay discreet and starts asking for more. The film portrays the mistress as mentally unstable and dangerous. The film ends with the wife killing the mistress and family values restored. The husband’s betrayal is never really an issue. I’m ashamed to admit that when I first watched it, some time in the 1990s, I sided completely with the poor husband, hounded by a woman the film had led me to see as a witch.

Both ideas have their roots in Genesis, where Adam, asked why he ate the forbidden fruit, answers that the woman gave it to him. It is easy scapegoating: powerful men keep their reputation for rationality while some snake of a woman takes the blame for their lapses of judgment. There’s a paradox that I find difficult to swallow. For centuries, women were judged too foolish to vote or hold office, yet the same culture credits them with enough cunning to bring down kings and CEOs. Men, meanwhile, are supposed to be fit to lead nations and markets while being too weak to resist the first temptation that comes their way.

Deadly clichés

The term comes from spy fiction: dictionaries credit John le Carré with popularizing “honey trap” in Tinker Tailor Soldier Spy in 1974. Espionage has long been a man’s world, and in its novels a female spy is very often a seductress. Real Cold War espionage looked quite different, since East Germany’s best-known seduction program used men, so-called “Romeo” agents sent west to seduce secretaries with access to government files. 

In the Nvidia story, much of the coverage of Alice Huang also draws on images Western culture has attached to Asian women for more than a century: the mysterious seductress, the Dragon Lady. In the U.S., they go back at least to the Page Act of 1875, the country’s first restrictive federal immigration law, which was enforced on the presumption that Chinese women were prostitutes. Any woman arriving from Asia had to prove she wasn’t one.

In March 2021, a white man killed eight people at three Atlanta-area spas, six of them Asian women, and police relayed his description of his victims as “sources of temptation.” Asian women end up being blamed for men’s desire. And Atlanta showed that this cliché can be deadly. Coverage casting Alice Huang as a seductress is anything but harmless.

Nobody trapped the CEO

A CEO’s sexual relationship with someone working for one of his company’s biggest customers is a clear conflict of interest. Alice Huang’s employer, Megaspeed, is said to have bought about $4.8 billion of Nvidia hardware over the past two years. In most companies, a sales manager sleeping with a key client’s buyer would be expected to disclose it and hand the account to a colleague. Nobody would describe that sales manager as “trapped,” and the CEO of one of the world’s most valuable companies, who has far more power, deserves no more indulgence.

The phrase also shrinks Alice Huang: if she is only “honey,” she cannot be much of a professional. Yet she ran a business that bought chips at enormous scale, and if the allegations hold, she will have to answer for export violations and espionage, serious crimes committed with the very agency the word denies her.

Call it what it is

The honey trap trope has consequences in the workplace. In 2019, a LeanIn.Org and SurveyMonkey survey found that 60% of male managers in the U.S. were uncomfortable mentoring, working one-on-one with, or socializing with a woman, and 36% of men said they had avoided doing so for fear of how it would look. A woman alone with a powerful man is seen as a danger to him, even though she is far more likely to be the one harassed. She also pays in lost dinners, business trips and sponsorship, the informal moments that decide who gets promoted.

Other words would be more honest. A CEO had an affair with an executive at one of his largest customers, a woman U.S. officials reportedly believe was working for Chinese intelligence, and his board is now examining whether the relationship compromised the company. It is less catchy than “honey trap,” but it treats both people as adults responsible for their choices. Isn’t that the standard any employee would expect?

Burnout is a huge workplace issue: the low engagement that comes with it costs every business affected a fortune. Worse yet, its effects on workers are deleterious to the point where it can literally lead to death.

Yet in my experience, most leader…

‘The Great Flattening’ has made companies faster and leaner. Nvidia’s scramble to replace its CEO shows what it can cost them.

Colette Kress never sought out to be a CFO. But after leading Nvidia’s finances for over a decade, she’s now tasked with replacing Jensen Huang.

Cisco’s focus on its people has helped earn it the No. 2 spot on the 2026 Fortune World’s Best Workplaces™ list.

Oxford Quantum Circuits closed the biggest Series C fund raise for a quantum computing company in Europe. Its CEO Gerald Mullally says the technology is finally moving to the commercialization phase.

A lot of AI advice suggests that you hand over work to a personal agent, like letting it triage your inbox or shop for you. That’s certainly appealing, since we’re all relentlessly busy and constantly fighting small tasks that take up our time. 

But every time you connect an AI tool to one of your accounts, you’re also answering a second question: What is this tool allowed to do on my behalf? 

For a lot of solopreneurs, that question gets answered by default. You click “allow” on the permissions screen, and the tool can now reach everything in that account. 

I use AI a lot in my business, but I’ve also drawn some very clear boundaries around what it can access.

Sort your tools by what AI can do in them

Every AI tool I use falls into one of three buckets.

Read but not edit. AI can pull information from the tool, but it can’t change anything. This works well for reference material you want AI to draw from without any risk of it being rewritten.

One example, for me, is pulling financial data from QuickBooks. I let Claude analyze some numbers every month, but I don’t want AI making changes to my transactions.

Full editing. AI can create and change things directly in these tools. This makes sense where a mistake is easy to spot and quick to undo.

I use Canva to plan a lot of my marketing material. Nothing is mission-critical, and the full ability to make changes gives me a lot of capabilities to use Claude with Canva.

Off-limits. AI can’t access the tool at all. 

One thing I don’t let AI touch at all is my bank accounts. JP Morgan Chase reported that people were handing over their banking credentials to AI agents like Muse. That’s outside of my comfort zone. 

Workarounds can ease your comfort with AI access 

Some of my most useful setups give AI partial access to a tool. 

Secondary accounts. You might give AI its own limited version of a tool and keep it out of your main account.

I do this with email. I have a secondary email inbox for inbound inquiries to my website. AI can sort and draft replies without touching my main inbox.

Limited permissions

Some tools will let you scope AI to certain projects. I use a database tool called Airtable to manage information for my solo business. Claude can only access certain parts of Airtable, so it can’t inadvertently cause an issue in a tool that I rely on heavily. 

“Supervised” tasks. AI does the work, and nothing moves forward until I approve it. With my email example, Claude can draft a reply to an email but can’t send it. The draft waits for me to review it.

Revisit the boundaries as the tools change

Your boundaries may change over time. AI tools keep adding new connections and capabilities. These changes may make you more or less comfortable with what they can do.

Or you might need to “use AI to fight AI.” I was resistant to letting AI access my email inbox, but the inbound inquiries I get from other people using AI themselves forced me to think differently. My solution was the separate email inbox. 

In the end, I can use AI as much as I do because I decided up front what it can access.

The 2026 midterms made one thing clear: Voters are “mad at” AI, as political strategist Rob Flaherty put it.

On the podcast Machine Gods, Flaherty notes that “the story people are saying to themselves is, ‘This thing is in my backyard. It is being used to train stuff that’s gonna take my job away, and I get no say over any of this, and I’m really mad about that.’” An NBC exit poll found that 56% of voters said that AI and data centers were a “crucial or major factor” in their vote for Congress.

It’s obvious that people hate data centers. But workers resent AI rollouts in their workplaces, too. Data tracked on Blood in the Machine shows office worker sentiment toward AI dropped from 81% positive in 2019 to just 43% positive (with 53% negative) by 2026, listing “force-fed tools” as a major grievance. A Gallup global study reported that 74% of Americans are worried about AI, driven by widespread anxiety over job obsolescence and automated layoffs.

These voters are American employees, many of whom are being asked by their employers to “transform” their work lives with AI. You might be thinking: How can someone who voted against fast-moving AI on Tuesday be open and willing to experiment with AI at scale in their work lives? The answer is they can’t. We have a negative bias towards uncertainty, and our AI future is the ultimate uncertainty. Nobody knows what’s going to happen. And as we saw in the election, voters said, “Slow down.”

American employers need to deal with this fact. That doesn’t mean voters—or employees—hate AI. For most of us, AI offers a fundamental paradox: There’s a lot we like about generative AI, how it helps our daily lives. And we are fearful of it. We have to hold that duality.

This matters deeply to any manager or leader who wants their company to adopt AI at scale. When people are fearful and conflicted, they want to find safety. This means we resist change, or we act out against it. “An April 2026 survey of 2,400 knowledge workers across the U.S., U.K., and Europe—including 1,200 C-suite executives—conducted by Writer and Workplace Intelligence found that 29% of employees admitted to actively sabotaging their company’s AI strategy, a figure that jumps to 44% among Gen Z workers,” writes Fortune. It’s very hard to be open, innovative, and experimental when you fear for your future.

Nothing about the election results says stop AI entirely. It says, “Pause and let’s please talk about this.” This has to happen at work, too. In this article, I share how to get started, with two key questions.

Why can’t people change?

I spent decades as a political and advocacy consultant, and clients always asked me the same thing: Why won’t people change their behavior? Why don’t they vote for the candidates who support their interests? Why don’t they exercise? Why won’t they support a policy that will save them money and improve their health? Here’s what I learned, and science backs me up: You can’t change behavior when people are stuck. So why do people get stuck?

• Change feels dangerous, and it’s safer to stay where you are.
• We’ve built an “immunity to change.”

In Robert Kegan and Lisa Lahey’s Immunity to Change framework, the idea is that when we can’t make a change we genuinely want, it’s usually not a willpower problem. It’s because some part of us is committed—quietly, even unconsciously—to the opposite. Kegan calls this a psychological immune system. It has one foot on the gas and one on the brake.

When change feels unsafe, our brain, nervous system and body will conspire to keep us in the status quo. After all, its job is to keep us safe. Often, our psychological immune system is run by anxiety. Anxiety is a response to threats we anticipate.

Change with AI came too fast, too soon, too much, which creates a threat response in humans. We freeze, fight, or flee. In his new book, Our Existential Advantage, Enrique Rubio argues that every major technological revolution teaches the same lesson: Things change. Historically, though, the change came slowly enough for society to adjust. The printing press spread knowledge across Europe and helped spark the Renaissance and the Enlightenment, but that took decades. AI is just as transformative, and it’s happening almost overnight. Because the change is so fast, the structures built on the old technology can’t keep up, and nor can we. As a result, many people despair that AI is something that’s happening to us, that it’s inevitable and we don’t have any agency over it.

We feel anxious about AI for very good reason. It threatens our safety, our livelihoods, our children’s futures, our values, and the meaning that we find in work. I’ve heard from employees who resent AI for siphoning off creativity and expertise, or think that it encourages sloppiness or laziness. We work for many different reasons. Although money is important, humans highly value agency at work, and we need meaning and purpose. When we give decisions to AI, what happens to core values like self-reliance, learning, or usefulness? The scholar Ziyaad Bhorat proffers three important abilities that individuals and communities have already identified as worth protecting from AI: work, knowledge and understanding, and decision-making. He writes that work is worth protecting, as is the right to keep our skills from atrophying and from being substituted by machines.

On Election Day, voters took back agency over AI. Their employers need to listen, and it starts with having a different kind of conversation about AI and work.

How to have the right AI conversations

Recently I was a guest on Changing Conversations, a podcast hosted by Katherine Cotter and Stig Albertsen, who lead organizational transformation work at the consultancy Implement. We discussed the opportunity leaders have right now to start conversations with their teams that understand the fundamental paradox of AI, and offer space and time to explore it. These conversations don’t need to have clear next steps and a defined outcome; they simply need to happen.

We are having technical conversations about AI: the opportunity, where to focus, how to roll out transformation and increase proficiency. There’s a real sense of FOMO with AI, a real sense of you’ve got to get with the program. A lot of people feel like admitting any sense of hesitancy or fear might make them weak. They worry that asking questions about AI rollouts might cause them to be viewed as un-innovative, which is ironic, because when you’re feeling stuck, you’re not as innovative. So the key is to get unstuck, and that means digging into our emotions around what AI means to us.

AI brings up a lot of emotions, and most of us are uncomfortable with emotions at work. That’s why this stuff is so important: When you don’t feel your emotions, you can’t make the best decisions. I want people to make good decisions. So I want them to sit for a minute and have this conversation, because when we’re able to name difficult emotions, understand them, and move through, we are able to create change, innovation, and great new things. The threat diminishes and we see a way forward. We regain our agency.

And it can start by asking two simple questions: “When I think of my craft, what makes me feel proud, fulfilled, purposeful?” and then “Does the implementation of AI change anything about this?”

I like to ask people to think of the parts of their craft and work that they really love to do, and that they feel mastery over. For most people, this is a good feeling. When we think about our craft, we feel proud, and we feel a tremendous sense of agency because we have worked hard to master something valuable. And when we then unpack how AI relates to our craft, it can bring up a lot of questions. It can remind us where AI is aligned with our values, and where it’s not. It can make us think about whether AI threatens our craft, or can enhance it, or both. This conversation opens up room to explore more deeply the question of what our AI future means for us.

I think AI gives us an opportunity to face what many of us see as an existential threat and to move through that. That’s how we gain control. Stig noted, “AI…holds an invitation to figure out what is the uniqueness of being human, or even what is my unique gift…. Part of what I do can be handled by other means and supporting tools and technology, and that’s an invitation for me to figure out: What is it that cannot be distributed to someone else, because only I can do that with that specific flavor?”

Questions to ask yourself and your team

I asked Stig: After all the years you’ve been doing this work, what’s the piece of your craft you love most, the thing you’re really good at? Try that yourself, and see what comes up. Sit with it; don’t rush to solve it!

Here are some more helpful questions to ask yourself and your team:

• What do I love about AI? What has surprised me?
• What parts of my job do I dislike and wish AI would take?
• What parts do I love and never want AI to touch? Why do they matter so much?
• What scares me about AI?
• What values are underneath all of that?

A few things I’ve learned from running these conversations: The leader goes first, and admits their own messiness. Expect the first few minutes to feel weird; people often look at me blankly until we get to their values or their kids, and then it clicks. Don’t attach a productivity goal. And make room for both/and, as the scholar Wendy Smith would put it. The most common thing I hear is some version of: I love it, it helps me so much, and I’m really worried about my kids.

People are relieved to learn they’re not the only ones holding both. The conversation is the most important part. There are no right or wrong answers, only the beginning of getting unstuck and moving through change.

Accounting was written off as too boring for millennials and easy for AI to kill—but now it could be a career path for millions of Gen Z NEETs.