The Day AI Got My Clients

By Sonal Nain

Newsletter Reporter/Producer

ShareNewsweek is a Trust Project member

In 2023, Edward Frank Morris, 33, based in Southampton, UK, made a decision most people would consider radical: he quit his job in copywriting. But Morris didn’t leave because he hated his job—he left because he saw the future arriving faster than anyone expected.

A 2025 Generative AI Susceptibility Index study found that by 2023-24 nearly all UK jobs exhibited some level of exposure to generative AI, with exposure levels in major cities—including London—reaching 35 to 36 percent of all roles. The shift isn’t just happening in the UK. In the U.S., for example, a Stanford study found that entry-level workers (aged 22–25) in roles most exposed to AI—like customer service, accounting and software development—have suffered a 13 percent employment drop since 2022.

Today, as CEO and Lead Prompt Engineer at Enigmatica, the UK’s leading Prompt Engineering and AI consultancy, Morris told Newsweek what drove him to “self-fire” and what it means for the broader workforce.

...

Like many copywriters before the generative AI boom, he had experimented with early “text spinners,” tools that rewrote or expanded existing content in seconds. “Back then it was Jarvis—now Jasper AI,” he recalls. “You’d feed it a brief, and it gave you content in minutes.” Morris had two crucial advantages: a long-standing interest in futurism and AI, and early awareness that Elon Musk’s side project, a small research lab called OpenAI, was quietly developing something far more powerful.

During the GPT-2 era, Morris signed up for beta access. Then, in November 2022, everything changed. He gained early access to ChatGPT or GPT-3 and instantly sensed its potential. Like many, he first used it playfully: funny emails, quirky blog posts, the odd message drafted for amusement. But the more he explored, the deeper the rabbit hole went.

“When I went on Twitter and clicked the ChatGPT tag, it was flooded,” he says. “People were building crypto bots, coding full programs from a sentence, getting ChatGPT to roleplay historical figures— even doing questionable, illegal stuff.”

Morris began spending hours experimenting: rewriting prompts, adjusting verbs and adjectives, trimming unnecessary words—effectively translating his human communication skills into AI communication skills. Over time, he realised that with the right prompts, ChatGPT’s writing was already on par with the average copywriter. “The writing was on the wall,” he says. “It was only a matter of time before my clients went to AI instead of me.”

“There have been huge leaps between ChatGPT-3 and ChatGPT-5.1,” he says. “Especially when you know how to prompt it.”

Today, he looks at Reddit and sees countless stories of writers being laid off because of AI—or accused of secretly using it. “Back in 2022 and early 2023, during the AI rush, people wanted to dig their heels in. They’d say, ‘No, I’m better than AI!’ and refuse to adapt. Where many of those writers are today, I honestly don’t know.” He adds: “Writers aren’t obsolete—far from it. But it’s harder than ever to thrive as a writer, copywriter, or content creator.”

Morris isn’t alone in witnessing the upheaval. Abi Hill, founder of Just Starting Out, a platform for emerging self-employed professionals, notes a growing trend: “People are turning to self-employment after roles were reduced, restructured, or replaced due to AI.” She observes that admin and entry-level creative roles are hit hardest—tasks like copywriting, scheduling, transcription and basic design are increasingly automated. Freelancers are also feeling the pinch, with clients using AI to draft content or handle design tasks.

The human cost is profound. Hill recounts one proofreader in her fifties, laid off from a national magazine after thirteen years due to “replacement technology,” whose career disruption contributed to a marital breakdown. Many stepping into self-employment do so reactively, unprepared for pricing, business insurance, or client acquisition—driven by necessity rather than choice.

Jenn Castro, 33, based in London, was a tenancy Progressor until the end of October 2025. “Everything was automated to the point that there simply was no need for my role,” she explains. Tenants could interact directly with AI bots or request customer support only when necessary. “Had the company not chosen to onboard this software—which I spearheaded—my role would have remained vital.”

With a shrinking lettings portfolio and the urgent need to cut costs, her position wasn’t deemed worth saving. The irony, Castro notes, was that during a two-week absence, six different colleagues had to cover elements of her work—and the same tasks continued to fall across multiple desks after her departure. With less than six weeks to train her co-workers before her final day.

“The automation sped up portions of my role, but they didn’t account for all the additional tasks I managed, the many other pies my hands were in,” she says. Castro highlights a recurring issue in AI-driven cost-cutting: even partially motivated by savings, decisions to eliminate human roles can have unintended consequences, creating inefficiencies and errors that ultimately offset initial gains.

Not every expert sees AI as a job killer. Aran Macfarlane, Senior Product Manager at TargetJobs, told Newsweek that AI is “genuinely transformative for certain tasks—information retrieval, data exploration, content drafting, pattern recognition and technical problem-solving.” But rather than replacing staff, it redistributes workloads: “AI handles repetitive tasks so you can focus on the interesting bits,” he explains.

Decisions about automation, Macfarlane notes, must account for resources, data security, and accuracy. Autonomous agents are still inconsistent, stating “OpenAI's agent has success rates of 38-87 percent depending on task complexity, so it's failing regularly.” AI still hallucinates, makes calculation errors, and struggles without human oversight, which can disrupt workflows and lead to unintended consequences.

Dan Bruce, an AI consultant at PressReacher, warns that companies rushing to replace employees often underestimate the value of human skills. “Critical thinking, perspective, and good judgment can’t be replaced by an algorithm,” he says. Businesses expecting instant efficiency find that small AI mistakes can snowball into larger problems. Hidden costs—software, integration, training, and consulting—often offset perceived savings.

Even in industries where AI seems strongest, humans remain irreplaceable. Tradespeople, creative professionals, and roles requiring real-world judgment resist automation. Some companies have even rehired staff they let go, creating hybrid roles blending human oversight with AI capabilities.

Many companies that rushed to automate have since expressed regret. According to a 2025 Orgvue survey of C-suite and senior decision makers, 39 percent made redundancies because of AI, and 55 percent of those now admit they made the wrong calls.

Even at a major fintech firm, the backlash has been real as Klarna laid off around 700 staff in its AI push, only to later backtrack and begin rehiring after admitting that the strategy hurt service quality

For repetitive administrative work, automation can free employees to focus on creativity, problem-solving and human connection. But over-reliance on AI risks moral, operational and strategic pitfalls.

“AI should allow us to reclaim our time and creativity while machines handle the monotony,” Morris says. Bruce adds a cautionary note for leaders: “Before cutting jobs, ask yourself—do we understand AI’s limits? Who manages its outputs? Which skills do we risk losing?”

The workforce is in flux. Some, like Morris, are preemptively pivoting to AI-centric careers. Others are navigating forced self-employment. Some companies are learning the hard way that replacing humans outright can backfire.

In the end, the future of work isn’t about AI versus humans. It’s about humans using AI wisely, carving out space for the uniquely human work that machines cannot replicate—and sometimes, knowing when to quit before the robots do it for you.

Request Reprint & LicensingSubmit CorrectionView Editorial & AI Guidelines