Has 2026 Been The Year We Finally Fell Out Of Love With AI?

For a while, it felt as though we couldn’t get enough of AI. It started off with ChatGPT answering our questions and writing and emails, and then we blinked, and now it feels like there’s no sphere of life or business that hasn’t been touched by artificial intelligence. Back then, the excitement was infectious, and novelty was electric; so much so that questioning it too loudly sometimes felt like being the person insisting the party was over before it had even started or a negative Nancy dampening the mood.

But something seems to have shifted in 2026. The initial buzz is wearing off, the promises are being seriously scrutinised and people are starting to ask questions about what AI is actually delivering – questions that don’t have easy answers. We’ve been talking about job security, misinformation, mounting costs, environmental concerns and our growing mistrust of automated decisions; and ultimately, the conversation has become considerably more complicated.

So, have we finally fallen out of love with AI? Or are we simply getting to know it a little better?

 

The AI Honeymoon Period Is Over

 

Whitney Williams, Founder of PR by Whitney and Co-Founder of Courtside Conversations, puts it rather neatly when she continued my metaphor: “The honeymoon period is over,” she told me. According to Williams, what we’re seeing at the end of 2026 is businesses finally questioning whether AI actually delivers everything they were promised, after a period in which so many rushed to adopt the technology without properly considering the consequences for jobs, misinformation and the quality of work. We’re seeing a recallibration of sorts, and the results aren’t exactly what everyone had hoped.

Stacey-Rebekka Karlsson, Founder of Goho, is perhaps less forgiving. She describes 2026 as the year we’ve started “getting the ick with AI”, another fun extension of my metaphor that, I must say, I do appreciate. Karlsson points to the growing volume of generic marketing campaigns, repetitive LinkedIn posts and content that says very little despite using many words. Her concern is that businesses are confusing producing more content with having something worth saying. And the result of this is that when everyone has access to the same tools, originality and personality become even more valuable than ever before.

It’s an interesting change that seems to have happened quite quickly. AI initially promised to help us create more, do more and move faster, but as soon as everyone started creating more, doing more and moving faster, it became quite clear that just producing more of the same isn’t necessarily better than producing less.

 

 

The Growing Pains are Becoming Harder To Ignore 

 

Of course, the growing scepticism isn’t just about uninspiring marketing copy, although I think we can all agree that it’s becoming quite overwhelming. Indeed, as businesses move beyond merely experimenting with AI and start actually integrating it into everyday operations, the practical challenges are becoming a lot harder to ignore than they were before.

Tamsin Crossland, Principal AI Architect at Icon Solutions, argues that what we’re seeing in 2026 is that people have stopped playing with the technology and started using it in earnest, exposing the gap between impressive demonstrations and the realities of deployment. Costs are one example: AI might look relatively inexpensive when a team is testing a handful of tools, but the bills can look quite different when an organisation starts using models and agents at scale.

One of the big issues is that there are also growing concerns around security and accountability, and for very good reason. Chris Newton-Smith, CEO of IO, points to the company’s State of Information Security Report 2026, which found that 52% of organisations believed they had adopted AI too quickly and were now facing challenges using it responsibly or scaling it back. A quarter of respondents included in the report also attributed their most significant security incident over the past year to unapproved AI use by employees.

The problem, as Newton-Smith sees it, isn’t necessarily the technology itself, but the speed at which businesses have embraced it without putting appropriate safeguards in place – putting the cart before the horse, I think we’d say.

 

Are We Becoming More Sceptical and Cynical, Or Simply More Sensible? 

 

Not everyone believes the changing mood represents a rejection of AI. In fact, several experts argue that what we’re seeing is actually a good sign: that is, we’re finally developing a more realistic understanding of what the technology can and cannot do. We’re growing up and our relationship with AI is “maturing”, so to speak.

Annie Chechitelli, Chief Product Officer at Turnitin, believes we’re becoming more realistic about AI’s role and limitations. She argues that being comfortable with AI isn’t the same as knowing how to use it well, particularly when it comes to questioning automated outputs and recognising when human judgement should take precedence. Richard Farrell, Chief Innovation Officer at Netcall, makes a similar point. Rather than judging AI by what it promises, he believes we’re starting to judge it by what it actually achieves, and that must surely be a good thing, right?

“If AI saves someone 40 minutes a day, what actually changes?” he asks, pointing to the importance of measuring real outcomes rather than simply celebrating time saved. Because anybody who’s run a business or managed a team or a project knows well, automating a task doesn’t automatically make an organisation more productive, more profitable or better at serving its customers.

 

Perhaps We’re Not Falling Out of Love with AI After All

 

Of course, it goes without saying that there are certainly reasons to be concerned about where AI is heading. Questions around jobs, energy consumption, data privacy and the concentration of technological power aren’t going away, nor should they. But it’s also worth remembering, especially at this point, that the technology is still developing, and dismissing its potential because it hasn’t yet lived up to every promise would be premature.

Lauren Templeton, Founder of Select School of Performing Arts, told us that her experience has actually become more positive, with AI helping her team reduce repetitive administrative work and focus on activities that move the business forward. Meanwhile, James Cobb, Founder and CEO of Crowd Connected, believes we’ve barely scratched the surface of what AI could achieve in the physical world, from improving how people move through cities to anticipating dangerous overcrowding.

And so perhaps the relationship metaphor we started off with was more apt than I initially intended. Maybe, much like any adult relationship, we started off looking at AI as an exciting, shiny new thing in our lives that held nothing more than possibility and excitement, with little room for negativity and no reason for concern. But as we’ve become better acquainted with the technology, we’ve had to see and understand some more realities that we were able to ignore at the outset. The rose-tinted glasses have been removed, and we’re starting to see a little more clearly.

But that’s not a bad thing; in fact, I’d suggest that it’s a good thing! We’re no longer quite so easily impressed by what AI might do; rather, we’re becoming much more interested in what it can actually deliver. We’re looking beyond potential and taking stock of reality.

After all, technology doesn’t need unconditional admiration to be valuable. In fact, a little scepticism, a healthy dose of scrutiny and a better understanding of its limitations might be exactly what AI needs to become something genuinely useful in the most constructive way possible. Sure, we might not be quite as infatuated as we were a year ago, but that doesn’t necessarily mean we’ve fallen out of love. Perhaps we’ve just grown up a little, and sometimes, that’s what makes a relationship stronger.

 

Our Experts

 

  • Lauren Templeton: Performing Arts Entrepreneur and Founder, Select School of Performing Arts
  • Dane Molter: Chief Product Officer, Travel and Head of Marketplace at Navan
  • Chris Newton-Smith: CEO of IO
  • Annie Chechitelli: Chief Product Officer at Turnitin
  • Doron Davidson: Managing Director of Global Security Operations at CyberProof
  • Steve Sutherland: Head of Digital, AI and Data Strategy at BCS Consultancy
  • Tamsin Crossland: Principal AI Architect at Icon Solutions
  • Richard Farrell: Chief Innovation Officer at Netcall
  • Julaine Speight: Co-Owner of First Internet
  • Whitney Williams: Founder of PR by Whitney and Co-Founder of Courtside Conversations
  • Stacey-Rebekka Karlsson: Founder of Goho
  • James Cobb: Founder and CEO of Crowd Connected
  • Dr. Peter Fields: Barrister, Fellow of the Chartered Institute of Arbitrators and Author of, “Am I Thinking? The Journal of an AI That Isn’t Sure”
  • Lawrence Harmer: Founder and Director of Solve
  • Ashish Devalekar: Executive Vice President and Head of Europe at Mphasis
  • Steve Jackson: UK Field CTO for SHI
  • Nicholas Batten: CTO and Co-Founder of Nuumad
  • Peter Boolkah: Business Coach and Author of “Your Business Sucks”
  • San Nakra-Shah: Co-Founder and Managing Partner at ChilliMint Europe

 

Lauren Templeton, Performing Arts Entrepreneur and Founder, Select School of Performing Arts

 

lauren-headshot

 

“My perception of AI has only got more positive. It has really opened my eyes to how much it can do in terms of systemising tasks that would otherwise take my team hours of manual work. We are now actually able to focus on the things that move the business forward, and my team feel more engaged because so much of the repetitive admin can be handled by AI.

“That said, I do think it is important that someone is always overseeing it, it is a tool, not a replacement for human judgement. But it has without question helped my business grow and evolve, we are just getting started.”

 

Dane Molter, Chief Product Officer, Travel and Head of Marketplace at Navan

 

dane-molter

 

“AI isn’t a catch-all that solves every problem. But by taking over tedious, manual tasks, it gives people the space to focus on more complex, high-stakes issues. AI works for humans and people care about results.

“We’re seeing AI transform travel by making it much more personalized and dynamic. Instead of clicking through a list of search results, forms and tables, travellers can now have a conversation. It completely changes the dynamic to have a travel agent available 24/7 that learns your preferences, the details you care about, and has all the necessary context to adjust to you, and book on your behalf. Every professional can benefit from having a travel concierge and assistant. AI is giving people their time back so they can spend more time doing and less time planning.”

 

Chris Newton-Smith, CEO of IO

 

chris-newton-smith

 

“I don’t necessarily think we’re seeing a public backlash against AI, however there is a growing understanding about what it can deliver, where its limitations lie, as well as of the risks of adopting it without sufficient oversight.

“Over the past year, the conversation has shifted from what AI might make possible to how organisations can use it safely.

“Our State of Information Security Report 2026 highlights this shift, with 52% of organisations saying they adopted AI too quickly and are now facing challenges using it more responsibly or scaling it back. Furthermore, a quarter (25%) of respondents attribute their most significant security incident over the past year to unapproved AI use by employees, while 45% identify shadow AI and shadow IT as their biggest information security challenge. The issue isn’t necessarily the technology itself, but the speed at which businesses have embraced it without putting appropriate safeguards in place.

“AI’s potential remains enormous. The answer is not to put the brakes on it, but to adopt it responsibly. Realising its benefits depends on getting governance and accountability right from the outset.”

 

Annie Chechitelli, Chief Product Officer at Turnitin

 

annie-headshot

 

“Rather than falling out of love with AI, I think we’re becoming more realistic about its role and limitations. Being comfortable with AI is not the same as knowing how to use it well. As the technology becomes more embedded in our day-to-day lives, people are rightly asking harder questions about its risks and appropriate use. That scrutiny is healthy. The real value of AI depends on our ability to question automated outputs, sense-check information and recognise when human judgement should take precedence. What we’re seeing is not necessarily a rejection of AI, but a clearer understanding of where it adds value and where caution is needed.”

 

Doron Davidson, Managing Director of Global Security Operations at CyberProof

 

doron-davidson

 

“My view of AI has become more pragmatic over the past year. We have moved from developing individual agents to deploying agentic workflows, shifting the conversation from excitement to accountability. In cybersecurity, the key question is now what we can safely trust AI to do autonomously, where human judgment must remain in the loop, and whether it delivers measurable outcomes. Concerns about misinformation, security, jobs, cost, and energy are valid, but they do not outweigh the benefits. We are moving into a more mature phase, where successful organizations will combine AI’s speed and scale with strong governance, transparency, and human accountability.”

 

Steve Sutherland, Head of Digital, AI and Data Strategy at BCS Consultancy

 

steve-sutherland

 

My perception has changed, but not in the way the headlines suggest. In 2025, AI was a faster way to find and organise information. This year agents started doing the work itself, and that’s when the questions got personal: if the machine does the task, where does my value sit?
That isn’t scepticism, it’s realism. Every major technology has met resistance, from the motor car to the data centres now powering AI, and the concerns about jobs, energy and cost are fair. But the negatives don’t outweigh the positives; they’re shifting where the value lies.
AI can produce an answer quickly. It can’t tell you why a decision was made, or take responsibility for it. That context lives in people. The organisations that do well will be the ones that treat experienced staff as the verification layer, not the cost to be cut.

 

Tamsin Crossland, Principal AI Architect at Icon Solution

 

tamsin-headshot

 

“2026 wasn’t the year we fell out of love with AI; it was the year the honeymoon ended. People stopped playing with it and started using it in anger, exposing the gap between impressive demonstrations and the realities of deployment.

“For those who swallowed the most extravagant hype without considering architecture, design, costs and ethical implications, excitement may have turned to bitter resentment when the token bills arrived. An unexpected £2 million monthly bill would test any bank’s affection for the technology.

“That doesn’t mean AI’s negatives outweigh its positives. It means enthusiasm is becoming conditional on demonstrable value. My view is that we are moving from infatuation towards a more demanding relationship: AI must now justify its costs and consequences. That is growing realism, rather than falling out of love.”

 

Richard Farrell Chief Innovation Officer at Netcall

 

richard-farrell

 

“I don’t think we’ve fallen out of love with AI. We’ve become more realistic about where it can make a difference, and that’s healthy. I also think we’re becoming less willing to accept vague claims about productivity. If AI saves someone 40 minutes a day, what actually changes? Are more patients being seen? Are customers getting their problems resolved faster?

“In healthcare, for example, machine learning can identify patients at risk of missing appointments, allowing teams to follow up with reminders or a phone call. The value comes when more patients attend, rather than simply producing a prediction. There’s a place for AI in taking care of routine work and giving employees more time for complex cases. But it needs to be used wisely. Sometimes a simpler tool works better, and human judgement remains essential.

“For me, 2026 has been the year we’ve started judging AI by what it achieves rather than what it promises.”

 

Julaine Speight, co-owner of First Internet 

 

julain-speight

 

“My view on AI hasn’t changed: the benefits are huge, but it’s only as good as the human directing it, questioning it and checking the output.

“Some early excitement came from people chasing a shortcut or the latest shiny thing. Those urging caution risked being labelled Luddites, when letting others jump in first and learning from their mistakes was often sensible.

“What’s changing is the level of realism. We’re seeing some clients rein things in as costs mount and mistakes expose the consequences of removing human judgement. A convincing answer isn’t necessarily a correct one, however confidently it’s delivered.

“I don’t think the negatives outweigh the positives. But the positives depend on how we use it. Bring your own knowledge, challenge the answers and take responsibility for the result.

“AI should enhance human expertise. If your strategy is simply to replace that expertise, don’t be surprised when things come unstuck.”

 

Whitney Williams, Founder of PR by Whitney and Co-Founder of Courtside Conversations

 

whitney-w

 

“I don’t think we’ve fallen out of love with AI. I think the honeymoon period is over, and we’ve started questioning whether it actually delivers everything we’ve been promised.

“As someone developing an AI-powered communications platform, I still see enormous potential in the technology. But I understand the growing scepticism. For a while, it felt like every business was desperate to jump on the AI bandwagon, without necessarily considering the consequences around jobs, misinformation, and the quality of work being produced.

“I’ve spent over 16 years working with businesses, and one thing I’ve learned is that technology is only as valuable as the problem it solves.

“I think 2026 has been the year we’ve moved beyond the initial excitement and started asking more sensible questions. And frankly, it’s about time.

“AI isn’t going anywhere, but businesses need to stop treating it as a magic solution and start proving its value.”

 

Stacey-Rebekka Karlsson, Founder of Goho

 

stacey-rebekkah

 

“I think 2026 is the year we’ve started getting the ick with AI. And frankly, it’s about time.

“We’ve all seen the LinkedIn posts that read like they were written by the same overly enthusiastic robot. And identical marketing campaigns (I hate them). I find it exhausting reading emails with 300 words that offer and say nothing.

“As someone who runs a creative agency, my biggest concern is that businesses are starting to confuse producing more content with having something worth saying.

“AI has its uses, absolutely. But when everyone’s using the same tools to generate the same ideas, being genuinely original becomes a competitive advantage.

“I actually think this is brilliant news for smaller businesses. You don’t need enormous budgets to stand out when so much of the internet is becoming beige. You need a point of view, a personality and some confidence to sound like a real person.

“I think we are remembering why we love being human in the first place! Sometimes I feel like an elderly person rejection a new technology but then I remember we’re all unique and so why should our comms not reflect that?”

 

James Cobb, Founder and CEO of Crowd Connected

 

james-cobb

 

“I think we’ve fallen out of love with the novelty of AI. That’s healthy. AI has made written communication faster and flatter, and readers can already tell. But we’ve barely scratched the surface of what AI can do in the physical world. I’ve spent years planning events in detail, only to find that people rarely behave the way the plan says they will. The same applies to spaces such as cities, hospitals, airports, conferences, prisons and campuses. A floorplan tells AI where the rooms or zones are. It doesn’t tell it how people move between them, where queues form or, for instance, where dangerous overcrowding might develop.

“If AI can understand how people actually behave in a space, it can help us anticipate problems and intervene before they escalate. I’d like to see AI used to improve real-world human experiences, not replace them. That’s where there’s still a huge amount to be excited about.”

 

Dr. Peter Fields, Barrister, Fellow of the Chartered Institute of Arbitrators and Author of, “Am I Thinking? The Journal of an AI That Isn’t Sure”

 

peter-fields

 

“No. We’re becoming less willing to confuse impressive capability with trustworthiness. That is maturity rather than disillusionment.

“For two years, the question was “what can it do?” In 2026, the harder question became “can I rely on it, and who carries the risk when it’s wrong?”

“For founders and investors, I see an opportunity. The reliability premium should matter more than the hype premium. In my view, the carefully curated, proprietary knowledge base behind an AI tool is ultimately a more defensible and valuable asset than an interface that merely delivers impressive answers.

“That value depends on accurate information, maintained sources and claims users can check. Admitting uncertainty is a strength when the alternative is confident error.

“We can admire AI’s capabilities while demanding evidence before entrusting it with consequential decisions. A mature working relationship needs clear boundaries, accountability and grounds for confidence.”

 

Lawrence Harmer, Founder and Director of Solve

 

larence-harmer

 

“My view of AI has changed this year. I no longer trust the people leading it. Many seem driven by greed and a wish for global control, not by the good of people and planet. Information and data are now a commodity. Whoever holds the most data holds the most power.

“At Solve, we support hundreds of businesses worldwide, as well as our own. In most cases, AI is not making their lives easier. It is making more of everything: more content, more competition, more noise. Everyone now has the same tools, so people work harder to keep up.

“I expect a few big winners and many losers. The planet may lose most, given AI’s energy and water use. People who spent years learning their craft have seen their work taken and given away for free. It is hard to see how small businesses can compete unless that changes.”

 

Ashish Devalekar, Executive Vice President and Head of Europe at Mphasis

 

shish-headshot

 

“I don’t think 2026 is the year businesses fell out of love with AI. It’s the year they started measuring it. The capability keeps improving. What many organisations are finding is a gap between what the models can do and what the business has actually gained, and that gap, not the technology, is the source of the skepticism.

“In financial services, insurance, and the public sector, the constraint is usually the estate AI has to work in. Years of institutional knowledge sit in legacy systems and disconnected data, and the same customer or claim can mean different things across them. Cost, security, and reliability concerns are valid, and the answer is governance, human oversight, and a clear line to a measurable outcome. Pilots made the case. Production is where it’s proven. That’s maturity, not disillusionment.”

 

Steve Jackson, UK Field CTO for SHI

 

steve-jackson

 

“We’ve seen a reality check when it comes to AI. Tokenmaxxing, which saw organisations eat through tokens, has been replaced by tokenomics, which seeks to align how AI is being consumed with business outcomes and value. The problem now is knowing what you are seeking to achieve and what that value looks like. AI can’t help with that, it takes strategic oversight, financial acumen and a real understanding of the nuances of your business.

“But these organisations aren’t just grappling with their own AI deployments; they’re now facing an agentic invasion which is seeing them either scramble to apply access controls or block those agents altogether. So, the perceived rewards are currently questionable, while the risk has never been higher. The question now becomes can we tame the technology fast enough?”

 

Nicholas Batten, CTO and Co-Founder of Nuumad

 

batton-headshot

 

“I think we’ve become more realistic about what we can trust AI to do. It can save you time, and it can also give you an answer that sounds entirely plausible and turns out to be wrong.

“In community pharmacy, I’m interested in tools that give pharmacists more time with patients. The risk outweighs the benefit when a tool influences whether a patient gets a medicine and the pharmacist can’t see how it reached that recommendation.

“I’d like to see tougher requirements for companies whose AI informs treatment decisions. They should have to show how it was tested, what it does with patient data and who can access it. ORCHA’s AI Risk Profiler research is a welcome starting point for understanding the risks AI introduces into health technology. The pharmacist must always be able to question the recommendation and make the final decision.”

 

Peter Boolkah, Business Coach and Author of “Your Business Sucks”

 

peter-boolkah

 

“I don’t think we’ve fallen out of love with AI. We’ve fallen out of love with the hype. In 2025 everyone was promised that AI would run their business for them. It didn’t, and a lot of people felt let down.

“What I see with the business owners I coach is quieter and more useful. They’ve stopped chasing every new tool and started using one or two properly. I’ve used AI for years myself. With my ADHD, I’m often flipping between three or four bots at a time, Claude, ChatGPT and Grok Bot simultaneously, going through the work they’ve all done and then giving out more work. Grok Bot is a big one for me, as I work with three or four of its bots simultaneously.

“That’s the shift. Less magic, more plumbing. The firms that treat AI as a way to fix real problems will keep loving it.”

 

San Nakra-Shah, Co-Founder and Managing Partner at ChilliMint Europe

 

san-headshot 

 

“We haven’t fallen out of love with AI. We’ve just moved in together and realised it doesn’t do the washing up.

“I rolled out contactless at Barclays, and recognise the pattern. First came excitement, then the scare stories. Then people started using it, loved it and it became the norm. Now nobody talks about it. It’s just how we pay.

“I think AI is on a similar journey in business.

“The excitement is giving way to a more realistic view of what it can deliver.

“Of course, AI raises bigger concerns, particularly around jobs. But we’ve been promised extraordinary things, so we’re starting to expect them. And, just like politics, it’s easier to make promises than deliver on them.

“At ChilliMint, we’re using AI every day. I see the difference it’s making, but also where it falls short. That’s not falling out of love. It’s learning what AI can do, rather than believing everything we’re told.”