By Stuart Kerr, Technology Correspondent, LiveAIWire
Between January and June 2025, companies reported over 77,000 AI-related layoffs in the United States. Freelancers in AI-exposed writing roles saw a 2 percent monthly job decline and a 5 percent monthly earnings drop over the same period, according to labour market analysis compiled in the SEOScaleUp dataset. Writing jobs overall fell 30 percent since 2022. Software and web development jobs fell 21 percent. Engineering roles fell 10 percent. These are not projections. They are the present-tense consequences of AI adoption by organisations that previously employed or contracted workers for tasks AI now handles faster and cheaper.
Yet the same data contains a finding that complicates the displacement narrative: independent workers — freelancers, contractors, and self-employed professionals — are also the demographic most rapidly adapting to AI. The World Economic Forum’s Future of Jobs 2025 report identified independent workers as both the most exposed and the most adaptable category in the AI labour market transition, because their professional identity is not tied to any single employer, their skill development is self-directed rather than institutionally constrained, and their survival depends on responding to market changes faster than employed counterparts. The freelance economy is experiencing AI disruption at the most acute level and generating the most creative responses to it simultaneously.
Where the Displacement Is Concentrated
The categories of freelance work most directly displaced by AI share a structural characteristic: they involved producing standard-format content or code at volume, where quality within a defined range was the deliverable rather than distinctive individual voice or problem-solving. Generic blog content, product descriptions, SEO articles, standard marketing copy, basic logo design, routine transcription, and formulaic code generation are all categories where AI tools can now produce acceptable-quality output at near-zero marginal cost. The freelancers whose primary income came from high-volume, lower-differentiation work in those categories have faced the steepest decline.
The Stanford Digital Economy Lab’s analysis of ADP payroll data found that entry-level positions for workers aged 22 to 25 in highly AI-exposed occupations dropped approximately 13 to 16 percent between 2023 and 2025. For freelancers, the equivalent entry point — taking on lower-paid, formulaic work to build a client base and portfolio — has shrunk for the same reasons. New freelancers in AI-exposed categories face a market where the low-complexity entry-level work that would previously have provided their initial income has been displaced by AI tools that their potential clients are now using themselves.
The earnings compression in AI-exposed freelance categories is measurable and significant. The 5 percent monthly earnings drop in writing roles represents a compounding decline that reduces annual income by more than half over 18 months at the same trajectory. For freelancers without savings buffers, this rate of earnings compression leaves insufficient time to transition to new service offerings before financial pressure forces them out of self-employment entirely. The adaptation that the research identifies as possible is real — but it requires capital, time, and market access that are not equally distributed across the freelance population.
The Adaptation Patterns That Are Working
The freelancers generating consistent income in AI-disrupted markets share identifiable characteristics that distinguish them from those experiencing the steepest declines. The most consistent pattern is the combination of deep domain expertise with AI tool proficiency — using AI to expand service capability rather than to accelerate production of commoditised output. A freelance marketing consultant who uses AI for market research, competitive analysis, and content drafting while providing the strategic interpretation, client relationship management, and contextual judgement that AI cannot replicate is doing qualitatively different and more valuable work than one who uses AI to produce more content at lower rates.
The PwC 2026 Global AI Jobs Barometer’s finding that professionalised roles — those where AI handles routine components, freeing human practitioners for higher-complexity activity — are growing twice as fast as democratised roles with 42 percent higher wage growth applies directly to the freelance market. The freelancers who are thriving have professionalised their services: they offer AI-augmented professional judgement, not AI-assisted volume production. A freelance data analyst who uses AI to process and visualise data at speeds they could not previously achieve, while providing the interpretive insight and business context that makes that data actionable, commands rates that reflect their amplified capability. A freelance translator who uses AI to handle the initial translation of large documents while providing the nuanced cultural adaptation and subject-matter accuracy review that AI still misses consistently positions themselves as a high-value AI quality control professional rather than competing with AI on raw translation volume.
The New Freelance Categories Being Created
AI is creating freelance market categories that did not exist three years ago. Prompt design for enterprise AI deployments — the specialised work of constructing effective instructions for AI systems at scale — has evolved from a curiosity into a genuine billable skill. Creative prompt engineers who combine domain expertise, brand voice knowledge, and AI tool proficiency to generate campaigns and content at scale are finding active markets among organisations that have AI tool access but lack the expertise to use them effectively.
AI output review and quality control is another growing category. Organisations that have deployed AI for content generation, code production, or data analysis need humans with sufficient domain expertise to evaluate AI outputs against professional standards and catch errors that AI’s hallucination tendencies and domain gaps produce. This is a role that requires the domain expertise that previously qualified someone for the work AI is now generating, applied to evaluating that work rather than producing it. The irony is notable: the professional made redundant by AI as a producer is potentially valuable as an AI output reviewer, provided they maintain and apply their domain expertise rather than having allowed it to atrophy during a period of AI-assisted production.
AI training and implementation consulting for small and medium businesses is the freelance category with perhaps the broadest accessible market. The majority of small businesses want to use AI tools but lack the expertise to identify which tools are relevant, implement them in their specific context, train their team to use them effectively, and establish the governance processes that prevent AI errors from harming client relationships. A freelancer with practical AI implementation experience and strong communication skills can serve a market that is growing faster than any other category of business services need.
The Independence Advantage
Independent workers have structural advantages in the AI transition that their employed counterparts lack. They can change their service offerings without waiting for employer approval, without navigating organisational inertia, and without the political complexity of proposing changes to established workflows. A freelancer who recognises that AI has displaced the demand for their current services can reposition immediately — updating their portfolio, adjusting their marketing, and targeting different clients — without the institutional friction that makes equivalent transitions difficult for employees within organisations.
The access to AI tools that freelancers enjoy is also structurally advantageous. Employed workers use AI tools provided by their employer, within the governance frameworks their employer establishes, for the tasks their employer designates. Freelancers choose their own tools, establish their own governance, and apply AI to whatever tasks in their workflow generate the most value. The freelancer who has experimented across the full range of available AI tools and developed a personalised workflow that integrates the strongest tools for each specific task they perform is likely better positioned than the employee limited to the enterprise AI suite their company has procured.
What Determines the Outcome
The factor that most determines whether a freelancer thrives or suffers in the AI transition is the specificity of their domain expertise. Generic freelance services — writing, coding, design, research — in which clients primarily seek quantity and format rather than distinctive expertise and judgement face the most intense AI competition. Specialised freelance services — writing about regulatory compliance in pharmaceutical manufacturing, coding in a specific niche technology stack with domain-specific requirements, designing brand identities for specific industry contexts — face meaningfully less AI competition, because the domain knowledge required to produce genuinely useful work in those niches is not well-represented in AI training data and requires human expertise to evaluate and direct.
For readers following the freelance economy’s intersection with AI, LiveAIWire’s coverage of what the IMF’s AI exposure figures actually mean for workers and our analysis of who is being left behind in the AI economy provides the broader framework. Ford’s recent admission that it had to rehire 300 veteran engineers after AI quality checks fell short identifies a concrete, real-world case of the specific capabilities that anchor the most durable positions, whether freelance or employed.
The Platform Economy Dimension
For freelancers who work through digital labour platforms — Upwork, Fiverr, Freelancer, Toptal, and their sector-specific equivalents — the AI transition is happening simultaneously on the supply and demand sides. Platform clients are using AI to do more of the work they previously commissioned, reducing demand for certain categories of freelance work. Platform workers are using AI to produce work faster and at lower cost, increasing supply and driving down rates for commoditised services. The combination is compressing rates and reducing volume in precisely the middle segment of the platform market — the work that is complex enough to be worth commissioning but simple enough that AI-assisted production significantly reduces the human expertise required.
The ghost work dimension of AI creates a specific challenge for platform freelancers that is rarely discussed in the context of AI and employment. The data annotation, content moderation, and AI evaluation work that is the fastest-growing category of platform work is also the most precarious and the least well-compensated — the work that the research on ghost workers documents as typically paying below local living wages, without contracts, benefits, or recourse. Freelancers who transition from AI-displaced content creation to AI-supporting data work may find that the transition involves a significant downward shift in earnings and working conditions. The adaptation pattern that is working for higher-skilled freelancers — moving up the value chain to work that requires domain expertise AI cannot replicate — is not equally available to all freelancers, and the pattern that is most available to lower-skilled freelancers at the bottom of the market is moving toward the most precarious and exploitative segment of the AI economy.
The Structural Advantage That Does Not Equalise
The independence advantage that freelancers have in adapting to AI disruption is real for those who have the skills, the financial cushion, and the market access to leverage it. The distribution of those assets across the freelance population is not equal, and the AI transition is not equalising it. The freelancers generating the strongest outcomes from AI adaptation are disproportionately those who entered the transition with deep domain expertise, professional networks that provide access to high-value clients, and savings buffers that allow them to invest time in skill development without immediate income pressure. The freelancers most exposed to displacement with the least capacity to adapt are those whose work was most commoditised, whose professional networks are thinnest, and whose financial situations leave the least margin for transition time. The AI freelance story is simultaneously one of genuine opportunity for those well-positioned to seize it and genuine vulnerability for those who are not.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, emerging technology, and their impact on business, society, and everyday life. LiveAIWire publishes original AI journalism every weekday at liveaiwire.com.