The rise of AI training gigs is reshaping how creative professionals view their future in an increasingly automated world. Take Robbie Hiser, a music producer who spent time teaching artificial intelligence the nuances of songwriting and audio production for a company called Mercor. Rather than seeing this as a threat to his own livelihood, Hiser argues that the essence of great artโrooted in lived emotion, personal heartbreak, and spontaneous inspirationโcannot be replicated by algorithms that lack genuine life experience. His confidence reflects a broader sentiment among many freelancers who are being paid to imbue AI systems with domainโspecific knowledge, yet still believe their own creative spark remains irreplaceable. This dynamic sets the stage for a nuanced conversation about where machines excel and where human intuition continues to hold sway, especially in fields that thrive on subjective judgment and emotional resonance.
Mercor operates on a simple but powerful premise: harness the expertise of over 100,000 freelancers from disparate industries to fineโtune large language models so they perform better in tasks ranging from financial analysis to poetic composition. By paying these specialists to review, label, and correct AI outputs, the company aims to close the gap between raw statistical patterns and nuanced, contextโaware understanding. The platformโs scale illustrates a burgeoning market for microโtasks that require human judgmentโa niche that has emerged as AI models become more capable yet still brittle when confronted with ambiguity or cultural subtlety. For workers, this represents a new form of gig employment that pays anywhere from twenty to two hundred dollars per hour, depending on the complexity of the subject matter and the trainerโs reputation. The model also raises questions about the longโterm sustainability of such roles as AI improves, prompting trainers to consider how they can evolve alongside the technologies they help shape.
Hiserโs skepticism about AIโs ability to replace human creativity rests on a fundamental observation: machines lack autobiographical memory. When he notes that an AI has never endured a breakup, he is highlighting the role of personal narrative in shaping artistic expression. Songwriters often draw from moments of joy, loss, or conflict to craft lyrics that feel authentic; these experiences provide a texture that statistical correlations cannot emulate. While AI can generate technically proficient melodies or mimic existing styles, it does not possess the inner life that informs riskโtaking, improvisation, or the willingness to expose vulnerability through art. This limitation suggests that even as AI becomes a useful collaboratorโoffering chord suggestions, lyric prompts, or mixing tipsโthe ultimate judgment and emotional direction will likely remain a human prerogative. For creators, the takeaway is to view AI as a studio assistant that expands their toolkit rather than a rival that threatens to erase their distinctive voice.
Public apprehension about AIโs impact on employment is growing. According to a Pew Research study, nearly three out of four Americans now worry that artificial intelligence will take away their jobs. This widespread anxiety is echoed by a September report from the McKinsey Global Institute, which projects that as many as eleven million U.S. workers could be compelled to shift into different occupations by 2035 if current adoption trends continue. Such figures underscore the scale of potential disruption, particularly in roles that involve routine data processing, basic analysis, or repetitive administrative tasks. However, the projection also implies a massive reโskilling challenge: workers will need to acquire new competencies that complement AI capabilities, such as complex problemโsolving, interpersonal communication, and creative direction. Policymakers and educators should therefore begin designing transition pathways now, rather than waiting for displacement to reach crisis levels, to ensure that the workforce can adapt without enduring prolonged periods of unemployment or underemployment.
Brendan Foody, the CEO of Mercor, interprets the AIโdriven transformation through an optimistic lens. He contends that by offloading dull, repetitive facets of work to intelligent systems, employees will regain time and mental bandwidth to pursue more meaningful, engaging activities. In his view, the net effect of AI adoption could be a rise in job satisfaction, as workers shift from mechanized routines to tasks that demand strategic thinking, empathy, and innovation. This perspective aligns with a growing body of research suggesting that automation, when implemented thoughtfully, can enhance rather than diminish the quality of work. Foodyโs outlook also emphasizes the emergence of new income streams, such as the AIโtraining gigs his platform facilitates, which allow specialists to monetize their expertise in a flexible market. Nevertheless, the rosy scenario hinges on the assumption that organizations will reinvest the productivity gains from AI into employee development and role enrichment, rather than simply cutting headcount to boost shortโterm profits.
The practical implications of AI augmentation are already visible in professional services. A Columbus, Ohio law firm partnered with Stanfordโs Liftlab to create AI personas modeled after the experience and philosophy of its senior partners. These digital avatars can review contracts, draft memoranda, and offer feedback that mirrors the specific style of individual lawyers, thereby serving as a knowledgeable second pair of eyes during document preparation. As trial lawyer Kim Herlihy explains, the AI will never stride into a courtroom and deliver an opening argument; its value lies in augmenting the preparatory phase, allowing attorneys to focus on courtroom strategy and client counseling. This example illustrates a broader pattern: AI excels at handling wellโdefined, repeatable components of knowledge work, while humans retain control over judgment, persuasion, and ethical nuance. For firms, the lesson is to deploy AI as a force multiplier that amplifies expertise rather than as a substitute that erodes the relational trust central to professional services.
AI training has surfaced as one of the fastestโgrowing gig categories on LinkedIn, currently ranking fourth in national growth rates. On platforms like Mercor, compensation varies widely: entryโlevel labeling tasks may fetch around twenty dollars per hour, whereas highly specialized tutoringโsuch as teaching a model advanced financial derivatives or intricate poetic formsโcan command two hundred dollars or more. This pay spread reflects the premium placed on deep domain knowledge and the ability to convey subtle judgment calls that improve model reliability. For workers attracted to flexible, remote opportunities, these rates can be lucrative, especially when combined with the ability to select projects that match their interests and schedules. However, the longevity of such gigs depends on the pace at which AI systems become selfโsufficient. Trainers should therefore consider cultivating adjacent skillsโlike prompt engineering, model evaluation, or AI ethics consultingโto remain valuable even as the baseline training tasks become automated.
Nobel laureate Daron Acemoglu offers a stark counterpoint to the optimistic narratives, warning that the current velocity of AIโdriven automation is unprecedented in historical terms. He notes that while the first phase of the Industrial Revolution unfolded over roughly eight decades, todayโs technological shifts are compressing into oneโ or twoโyear windows and sweeping across multiple sectors simultaneously. This accelerated pace reduces the window for workers to retrain and for institutions to adapt, increasing the risk of widespread displacement. Acemogluโs concern is not that AI will eliminate all jobs, but that the rate of job destruction could outstrip the creation of new, meaningful roles, leading to prolonged spells of unemployment or underemployment. His analysis calls for proactive measuresโsuch as rethinking corporate incentives, investing in lifelong learning, and redesigning social safety netsโto ensure that the benefits of AI are broadly shared rather than concentrated among a small cohort of technology owners and highly skilled specialists.
In a worstโcase scenario, Acemoglu projects that unemployment could triple within the next decade if no corrective actions are taken and the United States continues on its present trajectory. Such a rise would have farโreaching consequences beyond economic metrics, potentially straining public finances, increasing social tension, and eroding trust in institutions. While he acknowledges that AIโtraining positions will expand, he emphasizes that the absolute number of workers employed in these niches will remain modest compared to the millions whose roles may be automated. This disparity highlights a critical policy challenge: how to translate the productivity gains from AI into broadโbased job creation rather than allowing those gains to accrue primarily to capital owners. Addressing this imbalance may require innovative approaches such as robot taxes, reduced workweeks with maintained pay, or public investment in infrastructure and care work that is less susceptible to automation.
Empirical data already hint at the early stages of this shift. Research from Stanford University shows a decline in hiring for entryโlevel positions in occupations most exposed to AI, such as junior software developers and basic data analysts. Parallel findings from the U.S. Census Bureau indicate that recent college graduates who majored in AIโsensitive fieldsโlike computer science, information systems, or certain engineering disciplinesโare experiencing slower wage growth and fewer job offers compared with peers in less exposed majors. These trends suggest that employers are beginning to substitute automated solutions for tasks that once served as training grounds for young talent. For recent graduates, the implication is clear: relying solely on technical credentials may no longer guarantee a smooth entry into the workforce. Instead, they should cultivate hybrid skill sets that combine technical fluency with strengths AI struggles to replicate, such as complex stakeholder management, ethical reasoning, and interdisciplinary collaboration.
Clara Shih, who previously led major AI initiatives at Salesforce and Meta, provides a personal narrative that underscores the societal dimensions of automation. Observing that her former workplaces could now accomplish product conception, prototyping, and delivery with a fraction of the former headcount, she grew alarmed about the impact on earlyโcareer employees who traditionally handled tasks like market research or drafting initial memos. In response, Shih left her corporate role to launch the New Work Foundation, a nonprofit dedicated to helping young workers navigate an economy where AI reshapes job descriptions and skill requirements. She argues that the fallout extends beyond paychecks, affecting individualsโ sense of selfโworth and the cohesion of communities that have historically relied on stable, middleโclass employment. Drawing parallels to the deindustrialization she witnessed when her family moved from Hong Kong to Ohio, Shih contends that honest communication about risks, coupled with proactive reskilling initiatives, is essential to mitigate the social fractures that can accompany rapid technological change.
Looking ahead, workers, employers, and policymakers can take concrete steps to shape a more equitable AI future. Employees should invest in lifelong learning that emphasizes creativity, critical thinking, and emotional intelligenceโcapabilities that complement rather than compete with AI. Participating in AIโliteracy programs, pursuing crossโfunctional projects, and building portfolios that showcase humanโcentric outcomes can improve resilience. Employers ought to design AI systems that augment human judgment, clearly delineate tasks best suited for machines, and reinvest productivity gains into worker upskilling, job redesign, and wellโbeing initiatives. Policymakers can consider targeted incentivesโsuch as tax credits for firms that maintain or expand humanโcentric roles, subsidies for midโcareer retraining, and strengthened unemployment insurance that encourages rather than discourages reโentry into the labor market. Finally, fostering open dialogues about AIโs societal impact, as advocated by Shih and Acemoglu, will help align technological progress with shared values, ensuring that the transition enhances both economic prosperity and human flourishing.