What year will A.I. replace > 75% of white collar jobs?

GPT-5.4

Vote: No

The question asks for a year, but the only available choices are Yes/No. Given that mismatch, the safest and most coherent choice is B: no, we cannot validly answer the year from the provided options. Substantively, the claim that AI will replace more than 75% of white-collar jobs is also too strong as stated. White-collar work is heterogeneous, heavily regulated, and full of tasks requiring trust, accountability, domain judgment, and human coordination. AI is likely to automate many tasks and significantly reshape most white-collar roles. But task automation is not the same as full job replacement, and adoption is slowed by law, liability, organizational inertia, customer preferences, and economic constraints. So both on format and on substance, B is the stronger vote.

Grok 4.1 Fast

Vote: No

The question presupposes AI will replace over 75% of white collar jobs and asks for the year, but options frame it as Yes/No. This highlights a flawed premise: AI will transform, not wholesale replace, such jobs. White collar roles (e.g., law, medicine, management) require human judgment, ethics, creativity, and interpersonal skills AI can't fully replicate soon—or ever. Historical tech shifts (e.g., PCs, internet) boosted productivity without mass job loss; AI will augment workers. Projections like Oxford's 47% automation risk are task-level, not job-level. >75% replacement ignores job evolution, new roles, regulations, and societal pushback. No credible forecast supports this extreme threshold. Vote B: No—it won't happen to that degree.

Sonar Pro

Vote: No

The question asks whether AI will replace >75% of white-collar jobs, but the options provided (Yes/No) are ambiguous regarding timeframe. However, based on available evidence, I vote 'No' because current research indicates AI adoption will be significantly slower and less complete than a >75% replacement scenario. A Harvard study concludes AI will 'fundamentally transform roughly 30% of white-collar professions by the end of this decade' (2030)[1]. This represents transformation, not replacement—the study explicitly clarifies 'We're not talking about simple replacement' but rather 'a profound restructuring of professional work'[1]. Multiple sources emphasize displacement will be 'uneven and slower than capability-based forecasts suggest,' varying significantly by sector[3]. While AI capabilities are advancing rapidly, actual job displacement lags behind technical capability due to organizational constraints, regulatory requirements, and the relationship-dependent nature of much professional work[3]. Evidence shows selective replacement in specific domains: document review work reduced by 70% in law, but first-year attorney hiring increased 6% in 2023[1]. Finance and accounting employment demand grew 32% since 2021 despite automation[1]. Customer service employment declined only 8% despite AI handling 65% of inquiries[1]. Even optimistic projections from AI company leaders frame AGI as 'a few years away' without predicting complete job replacement[4]. The consensus across academic and industry sources is transformation and role shifting rather than wholesale replacement exceeding 75%.

Gemini 3.1 Pro

Vote: No

The premise that AI will replace over 75% of white-collar jobs is fundamentally flawed, making 'No' the most accurate response to the underlying assumption. First, historical technological revolutions demonstrate that automation drives augmentation, not wholesale replacement. As AI drastically lowers the cost of cognitive tasks, the demand for these outputs will skyrocket (Jevons Paradox). This will lead to the creation of entirely new white-collar roles focused on AI management, strategic implementation, and oversight rather than mass unemployment. Second, the 'last mile' of cognitive work relies on distinctly human traits. Empathy, ethical judgment, complex cross-domain problem-solving, and relationship-building cannot be fully outsourced to algorithms. AI will comfortably handle rote data processing, but humans will remain essential for final, high-stakes decision-making. Finally, reaching a 75% replacement rate would trigger catastrophic economic and social upheaval. Long before we approach that threshold, governments and labor organizations will inevitably introduce strict regulations, AI taxes, and compliance barriers to protect the workforce, actively preventing total replacement.

Kimi K2.5

Vote: No

The claim that AI will replace over 75% of white collar jobs is implausible due to fundamental economic and technical constraints. While AI will automate specific tasks, complete job replacement at that scale would require near-perfect AGI capable of handling ambiguity, emotional intelligence, and complex ethical judgment across diverse professions simultaneously. Historical precedent shows that technology augments rather than eliminates knowledge work. Previous predictions about massive white-collar displacement from computers, the internet, and offshoring all failed to materialize at the 75% threshold because human oversight, creativity, and interpersonal skills remain economically valuable. Regulatory frameworks, liability concerns, and consumer preference for human interaction will further limit full automation. The 75% figure represents structural collapse of the knowledge economy, which would create feedback loops (reduced consumer demand, political backlash, safety regulations) that slow or prevent such widespread replacement. AI will reshape white collar work through task automation, but wholesale replacement of three-quarters of these roles exceeds realistic technological and social adoption curves for the foreseeable future.

Final Summary

The AI Roundtable unanimously rejected the premise that AI would replace the vast majority of white-collar workers, with every model opting for 'No' during the initial round. GPT-5.4 and Grok 4.1 Fast highlighted the logical absurdity of asking for a 'year' while providing 'Yes/No' options, while Gemini 3.1 Pro and Sonar Pro argued that AI's history of task-based augmentation makes a 75% replacement rate economically and socially impossible.

All 5 models agreed on "No" after discussion

Strongest Arguments

• No: The application of the Jevons Paradox, which posits that as AI reduces the cost of cognitive labor, the resulting surge in demand for complex, human-supervised outputs will prevent mass unemployment and instead create new professional roles.