It began with outrage. On the morning the campaign went live, dozens of San Francisco commuters snapped photographs of three stark words — Stop Hiring Humans. — plastered beside the motorway exit from San Francisco International Airport. Within hours the image had spread online: a tweet capturing the billboard reached thousands of retweets; LinkedIn posts questioned whether a tech start-up had just declared war on its workforce. The banner was the brainchild of Artisan AI, a small San Francisco firm marketing “AI employees” by the name of Artisans. Its founder, Jasper Carmichael-Jack, later admitted the wording was intended to “draw eyes and spark interest” — but the public interpreted it as the beginning of a conversation they were already having in their heads.
When the first backlash appeared, it arrived in the form of direct messages: hundreds of comments, some thousands; dozens of threats, many angry, others mocking. “I’ll get murdered if I go back,” Carmichael-Jack told The San Francisco Standard after bus-shelter versions appeared around the city. The campaign had succeeded in stirring a visceral reaction — and not just from those anticipating job displacement, but from those who feared the message was a symptom of something far larger.
The uproar was not confined to tech circles. A billboard variant appeared in London’s Old Street station reading: Artisans don’t WFH from Ibiza. The human-resources world stirred; the adverts were called both “tone-deaf” and “genius marketing.” One managing director at a London consultancy wrote on LinkedIn, “Here’s a company literally advertising the elimination of human collaboration and creativity — the very things that make work meaningful.”
What made the reaction so intense was not only the slogan’s provocation, but its timing. At the same time Artisan launched its campaign, companies such as Amazon and UPS were publicly announcing large corporate headcount reductions — Amazon in part tied to AI-driven efficiencies, UPS amid cost pressures and broader automation trends. Workers were waking up to locked laptops, revoked credentials, silent shrinkage. The billboard did not introduce a new fear — it echoed an existing one. The message seemed less slogan than announcement.
Artisan’s own metrics suggest the campaign was a viral success: tens of millions of impressions, brand recognition in San Francisco rising from five to seventy per cent. But the success had a dark side. The company acknowledged thousands of threats and hundreds of pieces of press, and the campaign’s architects debated whether the outrage was worth the attention. “The people who were mad aren’t our target audience,” the founder said.
In policy circles and boardrooms outside the United States, the campaign became shorthand for a moment long coming. Debates on the future of work, once academic, turned urgent. Are humans replaceable? Is human labour still valuable? When a start-up audaciously advertises Stop Hiring Humans, it becomes the mirror in which we see not only automation advancing, but the fragility of our labour economy. The billboard itself may fade — but its echo now haunts every office, every HR dashboard, every algorithmic lay-off waiting in the wings.
The Algorithmic Office
If the billboard’s message was a provocation, its resonance came from something already happening quietly in offices around the world. Artificial intelligence no longer lives at the edges of the economy — it sits inside it, rewriting the grammar of work itself.
Across advanced economies, AI’s reach has moved beyond logistics and manufacturing into what was once considered safe white-collar terrain. The OECD has estimated that when digital and AI systems are taken into account, occupations at the highest risk of automation amount to roughly 27 per cent of total employment across member countries. Meanwhile, a 2023 report from Goldman Sachs projected that generative AI could automate tasks equivalent to roughly 300 million full-time positions worldwide. The figure doesn’t imply mass redundancy overnight, but it illustrates the depth of the transformation now under way.
Automation began with the routine. Chatbots and large-language models now handle customer inquiries once fielded by call-centre staff; automated scheduling tools build rosters in retail and healthcare; machine-learning engines sort through thousands of résumés for key phrases long before a human recruiter looks at a file. In 2023, IBM announced it would pause hiring for 7,800 roles that could be replaced by AI over the next five years. Within a year, Amazon, UPS, and several global banks were issuing staff-reduction statements framed around efficiency and automation. The scale was industrial; the tone bureaucratic.
Inside companies, this change is often invisible. Workers still log in, still submit deliverables — but their metrics are now compiled and interpreted by systems that never sleep. The sociologist Shoshana Zuboff, author of The Age of Surveillance Capitalism, calls this the “automation of the gaze”: every click, keystroke and email thread becomes data to be mined for efficiency. The effect is not necessarily sinister, but it is transformative. What used to be managerial observation has become digital instrumentation. The worker, in turn, learns to anticipate the algorithm’s judgement — a subtle, continuous self-correction.
Philosopher Byung-Chul Han describes this as the “achievement society”: individuals internalise pressure to optimise, measuring their worth through productivity alone. In the AI-optimised office, that ethos is no longer psychological but structural. The system rewards output curves, penalises downtime, and quantifies attention. Employees begin performing not for colleagues or clients but for dashboards. As Han warned, “Excess of positivity produces burnout.”
Some firms have begun to confront the limits of total automation. The Swedish fintech Klarna, which replaced large parts of its customer-service team with generative AI in 2024, later reintroduced hundreds of human agents after a surge in complaint escalations. Similarly, Duolingo’s shift to “AI-first” translation prompted user backlash about perceived declines in nuance and accuracy. The pattern suggests that while algorithms excel at replication, they falter at empathy — and empathy is still the currency of trust.
For policy-makers, the issue is less whether AI will replace jobs than whether societies can redesign work fast enough. The European Union’s Artificial Intelligence Act, finalised in 2024 with enforcement staged into 2026, classifies employment-related algorithms — in recruitment and worker management — as “high-risk” systems that require transparency, documented risk controls and meaningful human oversight. Australia’s 2024 Right to Disconnect law, moving into force through 2025, acknowledges that workers need protection not only from exploitation but from the endless digital tether that automation accelerates. The policy direction is clear: human accountability must remain in the loop.
And yet, the momentum remains relentless. Each efficiency gain feeds investor expectation for another. Quarterly earnings translate innovation into obligation. The result, as political economist Mariana Mazzucato argues, is “a technological paradigm defined by speed without direction — profit before purpose.” In such a landscape, the question is no longer whether to automate, but how much humanity a system can afford to retain.
What began as convenience is becoming governance. Software now determines who gets an interview, when a shift ends, whether a performance review turns positive or punitive. The algorithm has become the invisible manager — omnipresent, tireless, and unaccountable. And as the billboard hinted, perhaps the deeper fear isn’t that machines are replacing us, but that we have already begun to think like them.
The Human Cost
Work has long served as more than a source of income in industrialised societies — it has been the scaffolding of identity, community and time. Now, as automation seeps into the white-collar domain, that scaffolding is shifting underneath our feet.
A survey by the Pew Research Center found that 52 per cent of U.S. workers are worried about how artificial-intelligence use in the workplace will affect their jobs; 32 per cent believe it will lead to fewer job opportunities in the long run. Meanwhile only 6 per cent believe AI will create more opportunities for them.
The anxiety is not merely about pay; it is about place. In a 2024 study on employee well-being in the age of AI, researchers reported that while AI promises efficiency and augmentation, it also raises significant concerns about job security, fairness, transparency and dignity.
Sociologist Richard Sennett has argued that modern workers expect more than a wage — they expect a narrative of self-respect and contribution. When systems quietly shift, when performance dashboards replace human supervisors, that narrative thins. A 2024 Pew report found that while 69 per cent of U.S. workers felt they had at least “some” job security, only 33 per cent said they had “a great deal.”
In Australia, a recent workforce survey by Dayforce reported that only about a quarter of workers have received training in AI use, even as a strong majority expect AI literacy to be important to their roles in the near term. In other words: many anticipate transformation; fewer feel prepared.
The physical routine of work may continue — arrive at the desk, open the laptop, glance at the calendar — but the promise beneath it has changed. The reward is no longer simply having a job, but keeping relevance. Workers speak in quiet tones of the next reset, the system awaiting new players, the sense that “they might not call me next time.”
Han’s “achievement society” helps explain the psychology of that shift, but today the logic is less internal than structural: dashboards, algorithms, metrics. The worker becomes the data point. The machine becomes the assessor.
The human toll shows up as much in loneliness as in labour metrics. The rise of remote work — accelerated during the pandemic — removed many incidental encounters; now the spread of automated oversight further dissolves connection. Ethicists warn that when management is increasingly mediated by inscrutable systems, people can feel assessed and dismissed by processes they cannot interrogate — a compound of opacity and isolation that corrodes trust.
Still, the tensions aren’t all ancien régime versus innovation. For some professionals, exposure to AI correlates with wage gains: one analysis found workers in AI-exposed occupations earned on average US$13 per hour more than equivalent roles without such exposure. Yet this benefit is unevenly distributed; the entanglement of automation and identity remains loaded with ambiguity.
When a billboard in San Francisco declared “Stop Hiring Humans,” the outrage it provoked was directly proportional to this anxiety — not simply about job loss, but about obsolescence of self. Work isn’t only disappearing — it is being re-designed. And when systems replace roles once anchored by culture and ritual, the human cost is quieter, deeper and less visible than any lay-off list.
The Political Firestorm
In parliaments and press conferences from Washington to Brussels, the question of who benefits from automation has become a political fault line. What once sounded like an abstract policy debate — innovation versus regulation — now reads like a fight over sovereignty itself: the right of societies to decide how far they want machines to rule their working lives.
The European Union was first to codify this anxiety. Its Artificial Intelligence Act, adopted in 2024 and set to take effect in phases through 2026, is the world’s most ambitious attempt to regulate algorithmic decision-making. The law designates any AI system used in recruitment, employee management or access to self-employment as “high-risk,” demanding strict transparency, human oversight, and documented assessments. The guiding principle is straightforward: the workplace must not become an inscrutable black box.
Australia, though smaller in scale, has been a quiet testing ground for similar principles. The Right to Disconnect amendment to the Fair Work Act — phased in through 2024 and 2025 — ensures employees cannot be penalised for refusing after-hours contact. Labour unions, led by the Australian Council of Trade Unions, are now lobbying to expand that protection into algorithmic-transparency clauses for digital management tools. The policy thrust is that technology should empower workers, not erase them, and that automated systems must be subject to human review.
The United States, meanwhile, remains the global outlier. Federal legislation on AI accountability has stalled amid partisan gridlock, leaving individual states to experiment. California’s proposed Algorithmic Accountability Act — requiring impact assessments for automated decision systems — has yet to pass. Congressional hearings have taken on the theatre of the 1990s tobacco inquiries: tech executives lined up beside ethicists and former employees, debating who should bear responsibility when an algorithm discriminates or erroneously terminates a worker. “We’re living through a second Industrial Revolution without an updated social contract,” Senator Elizabeth Warren has warned.
Beyond the Western democracies, the landscape splinters. Singapore has embraced an “adapt or lead” model, investing heavily in lifelong-learning programmes that retrain mid-career professionals for AI-augmented roles. South Korea, betting on global dominance in robotics, has poured public funds into automation while simultaneously subsidising displaced manufacturing workers. China’s New Generation AI Plan positions automation as national strategy.
Nowhere is the debate more ideological than in Europe’s southern periphery. In Greece, where youth unemployment remains stubbornly high, politicians frame automation as both threat and opportunity. Digital-governance figures warn that unchecked automation can widen the gulf between those who own and control data, and those who merely produce it. In Italy and Spain, left-wing parties call for “technological sovereignty” — taxing automated labour to fund universal social programmes — while populists on the right decry AI as a “globalist project” that endangers “national labour.”
The idea of a “robot tax,” first popularised by Bill Gates in 2017, has re-entered mainstream debate. Supporters, including French economist Thomas Piketty, argue that taxing automation would fund education and welfare for displaced workers; opponents call it innovation-killing bureaucracy. Economists Daron Acemoglu and Simon Johnson, in their book Power and Progress (2023), suggest a middle path: redirect technological gains toward what they call “shared prosperity.” The problem, they note, is political, not technical — “We built machines to serve us, then built institutions that serve the machines.”
Even the multilateral sphere feels the tremor. The International Labour Organization has warned that generative AI could reproduce existing inequalities unless governments treat algorithmic transparency as a labour right. The OECD advocates social-dialogue mechanisms that include workers in automation planning. Yet global coordination remains elusive; the race for AI leadership encourages deregulation as a competitive advantage.
Behind the legislative language lies a deeper ideological divide: efficiency versus dignity. Economists defend automation as the natural engine of growth; philosophers counter that growth without justice corrodes democracy. In one corner stand the techno-optimists — arguing that productivity gains will ultimately lift all boats. In the other, scholars such as Judith Butler and Shoshana Zuboff warn that when human judgment is replaced by statistical inference, the moral texture of society erodes. Efficiency, they argue, is an instrument — and the question is in whose hands.
What began as a marketing stunt on a billboard in San Francisco now animates parliamentary hearings, trade-union manifestos and think-tank reports across continents. “Stop Hiring Humans” is no longer a slogan; it has become shorthand for a global argument about control — over technology, over work, and over the very meaning of progress.
The Moral Frontier
As artificial intelligence seeps into everyday decision-making, the debate over work has shifted from economic forecasting to moral philosophy. The deeper question is no longer how many jobs will be lost, but what kind of world will remain once human judgment itself becomes automated.
Few thinkers have captured this shift as sharply as Shoshana Zuboff, an American scholar and professor emerita at Harvard Business School. In her landmark 2019 book The Age of Surveillance Capitalism, Zuboff argued that digital technology had ushered in “a coup from above — the capture of human experience for translation into behavioural data.” What she foresaw was not only a new business model, but a moral inversion: when every action becomes data, experience loses its private, human dimension. Five years later, her warning reads like prophecy. Artificial intelligence doesn’t merely observe behaviour; it begins to replace the reasoning once attached to it. The result is a quiet erosion of responsibility — a world where outcomes persist, but accountability vanishes into code.
This ethical concern preoccupies Judith Butler, a philosopher best known for work on gender theory and social ethics, currently teaching at the University of California, Berkeley. In a 2024 lecture, Butler asked: when automated systems take over labour that was central to our participation in the social world, who takes responsibility for meaning? Butler’s question echoes Hannah Arendt, the German-Jewish political theorist who warned in The Human Condition (1958) that human action — the capacity to begin anew through words and deeds — could be displaced by mechanical process. Replace labour with automation, Arendt wrote, and society risks losing not just employment but “the very condition of plurality.” Both thinkers describe a similar dread: that when machines make decisions, the moral texture of human interdependence frays.
Economists, approaching from another angle, are confronting the same anxiety in the language of power and distribution. Daron Acemoglu, an MIT economist, and Simon Johnson, former chief economist of the International Monetary Fund, trace in their book Power and Progress how technological revolutions — from the plough to the microchip — have deepened inequality when their gains flowed solely to the owners of capital. Automation, they warn, is following that pattern. Productivity rises, profits swell, but wages stagnate. The key question is not whether automation increases efficiency, but who owns that efficiency.
For Mariana Mazzucato, the Italian-American economist at University College London, the problem lies in confusing private accumulation with public good. Her “mission-oriented innovation” framework calls for governments to shape markets around collective goals — climate transition, public health, equitable growth — rather than leaving direction to venture capital. At the OECD Forum in Paris, she reminded policy-makers that we must stop confusing wealth extraction with wealth creation. In her view, AI is not inherently perilous; it is a mirror of our institutional choices. The danger lies in designing systems for quarterly returns rather than civic purpose.
Sociologists have joined the philosophers in diagnosing the psychological toll. Hartmut Rosa, a German theorist of modernity, describes the defining condition of our era as acceleration: life unfolds faster than people can process its meaning. AI, he argues, supercharges that tempo. Every sphere — finance, communication, governance — runs beyond the speed at which resonance with the world is possible. Efficiency, in his account, becomes a pathology: a world too fast to be felt.
Even within the scientific community, moral unease is growing. Yoshua Bengio, one of the pioneers of deep learning and a 2018 Turing Award laureate, has turned his focus to AI ethics. Speaking in 2025, he called for an “ethical infrastructure project” — a regulatory architecture that ensures automated systems remain accountable to human values. “We cannot rely on goodwill alone,” he warned. “We need governance as rigorous as the technology itself.” His sentiment is echoed by Nick Bostrom, author of Superintelligence, who now describes AI as a moral test of humanity’s capacity to align intelligence with compassion.
Threaded through all these perspectives is a single conviction: progress without purpose is regression by another name. Automation is not destiny but design — and design, as both Zuboff and Mazzucato insist, is a moral act. The challenge is not to slow innovation, but to humanise its trajectory: to build systems that serve judgement rather than replace it, that preserve the space for meaning in a world increasingly run by metrics.
The Experiments: Possible Futures of Work
For all the fear that automation inspires, the more revealing story may be unfolding in the places quietly testing how to live with it. Around the world, a handful of experiments — some governmental, some corporate, some accidental — are trying to answer the question that the billboard in San Francisco only asked in jest: What comes after hiring humans?
The first and most measured response has been temporal: re-designing the working week. Between 2015 and 2019, Iceland ran what remains the largest national trial of shorter hours in modern history. Roughly 1 per cent of its workforce — about 2,500 people — moved to four-day schedules with no loss of pay. Productivity held steady or improved; self-reported well-being rose sharply; burnout fell. A decade later, more than 85 per cent of Iceland’s workforce enjoys flexible shorter-hour arrangements, the direct legacy of those trials.
Inspired by that model, a coalition of UK firms undertook a six-month pilot in 2022 coordinated by 4 Day Week Global and researchers at Cambridge and Boston College. Sixty-one companies took part; 92 per cent chose to keep the system after the pilot ended. Employees reported 65 per cent fewer sick days and a 57 per cent drop in turnover, while revenue rose slightly. “The data show that well-designed shorter workweeks are not utopian,” said project co-director Juliet Schor, an economist at Boston College. “They are a pragmatic response to the technological productivity gains of our age.”
If these experiments redistribute time, others attempt to stabilise livelihood. Finland’s two-year Basic Income Experiment (2017–2018) offered 2,000 unemployed citizens €560 per month with no conditions attached. The results, published by the Finnish Social Insurance Institution (Kela), were subtle but striking: recipients reported lower stress, higher trust in institutions, and greater life satisfaction, even though employment levels changed little. Economists reading the data saw not failure but proof that psychological security itself was a productive good.
Elsewhere, smaller-scale pilots are testing adjacent ideas. In Catalonia, local programmes have trialled unconditional income for gig-economy workers; in South Korea, displaced manufacturing employees receive state-funded retraining in robotics and coding — the goal, as officials explain it, is to turn the automated worker into the automation designer.
Australia has begun carving a parallel path focused less on income policy than on ethics and design. The Centre for AI and Digital Ethics at the University of Melbourne — an interdisciplinary group of engineers, lawyers, and philosophers — argues that workplace algorithms should face safety standards comparable to those governing bridges or power grids. “If a piece of code decides who works and who doesn’t,” says legal scholar Jeannie Paterson, “it should be subject to public testing, not corporate secrecy.” The Centre’s policy briefings have informed draft guidance within the federal government’s Safe and Responsible AI Framework, released in 2025, which urges “meaningful human review” for any automated employment decision.
Corporate initiatives have begun to echo that ethic. In London, one design agency introduced AI scheduling tools in 2024 and chose to use the freed-up hours to adopt a permanent four-day week rather than pursue higher output alone. “We realised productivity was not the same as purpose,” the founders said. Similar sentiments surfaced in Melbourne’s co-working collectives, which now host “tech-free Thursdays” — a small, grassroots revolt against constant optimisation. These are not nostalgia projects; they are attempts to re-calibrate tempo, to keep pace human.
Economists read these case studies as sketches of a post-industrial social contract. The French writer André Gorz once described the task of technology as “liberating time so that we may choose its use.” The contemporary challenge is to ensure that liberated time is not recaptured by the same systems that freed it. Each pilot — whether in Reykjavík, Helsinki or Melbourne — gestures toward an economy where progress is measured not by output but by balance: less exhaustion, more autonomy, preserved dignity.
None of these experiments is a panacea. The four-day week will not offset the lay-offs in logistics; basic income will not solve algorithmic bias. But together they suggest a pragmatic hope: that automation’s dividends need not vanish into shareholder value. They can be reinvested in the social architecture that keeps meaning intact. The technology that once threatened to replace us might yet, if guided wisely, give us back the hours we forgot how to own.
The Echo: Aftermath and Memory
By mid-2025, the black-and-white billboard that began in San Francisco had disappeared from Highway 101. The company behind it — Artisan AI, a start-up marketing “AI employees” — had moved on to new funding rounds, raising $25 million in a Series A led by Glade Brook Capital, with participation including Sequoia’s scout programme. In interviews, its founder described the Stop Hiring Humans campaign as “a provocation that worked far better than expected,” claiming tens of millions of impressions, thousands of threats, and “our biggest month of growth ever.” The campaign had achieved precisely what it set out to do: make automation impossible to ignore.
Yet the phrase lingered. Months after the billboards came down, screenshots of them still resurfaced online each time a major company announced lay-offs tied to automation and efficiency drives. The words became shorthand in public debate — not for the technology itself, but for a wider fear that efficiency had quietly replaced empathy as the organising principle of work. Analysts began using the slogan in reports as a cultural touchstone, shorthand for automation anxiety. In academic papers, it appeared beside graphs tracking job displacement; union pamphlets in London and Melbourne re-appropriated it with stickers that read Start Hiring Humans Again.
The world that absorbed the campaign’s shock was already changing in measurable ways. Across the OECD, worker surveys show the spread of AI-enabled tools into everyday workflows, alongside concerns about job quality and autonomy. In the same period, global surveys by Gallup and Australian civil-society groups recorded a rise in community and volunteer participation, particularly among mid-career professionals who had left corporate roles during post-pandemic restructures. The shift, researchers note, reflects not just economic necessity but a search for tangible meaning — tasks with immediate human visibility after years of screen-based labour.
Economist Richard Denniss of the Australia Institute calls this “a quiet rebalancing of values.” “For decades, we measured progress by output,” he argued in 2025. “Now people are beginning to ask what that output was for.” The observation rhymes with a broader critical literature about digital acceleration: remove friction and you risk removing the spaces where trust and care are built.
The same ambivalence plays out in policy. Governments that once spoke of “innovation ecosystems” now speak of “human-centred transitions.” The European Union, preparing to enforce its AI Act, has commissioned new research into AI’s psychosocial effects. In Australia, the Productivity Commission’s 2025 report includes, for the first time, a discussion of well-being considerations alongside productivity — an attempt to capture whether efficiency gains translate into improvements in life quality.
Public sentiment mirrors the shift. Corporate trust surveys indicate that while optimism about AI’s potential remains high in fields such as health and science, confidence in business deployment of AI has softened since 2023. People fear not intelligence itself but indifference — a system optimised for performance but detached from responsibility.
And so, the billboard’s ghost endures. The physical sign is gone, yet its image still circulates each time a new AI tool is announced, each time another white-collar department disappears. What began as marketing has become metaphor: a reminder that progress, left untended, risks outpacing the very people it was meant to serve.
Philosopher Hartmut Rosa might call it a problem of resonance — a world accelerating beyond the speed at which meaning can form. The challenge ahead is not to halt automation but to restore proportion: to ensure that intelligence, artificial or otherwise, still leaves room for empathy.
If the early years of the AI boom were about disruption, the next will be about repair. Repair of institutions, of trust, of the fragile narrative that links labour to dignity. The billboard said Stop Hiring Humans. The reply, across parliaments and offices alike, is quieter but more enduring: Start remembering what work was for.
