Ekkehard Ernst, chief macroeconomist at the International Labour Organization (ILO), spoke in Beijing on Tuesday.
Ernst warned about "algorithmic collusion" as a greater threat to employment than job displacement by AI.
He stated that the disruptive potential of AI on labor markets has been overestimated.
A study by American AI company Anthropic, released this month, was cited to support the claim of an "implementation gap" in AI adoption.
The study found that AI's real-world adoption lags due to regulatory hurdles, system integration complexities, and the need for human oversight.
AI is currently impacting specific sectors, such as software engineering, and entry-level roles.
Ernst noted that youth unemployment is primarily related to the current economic slowdown rather than AI.
The discussion took place during a public address in Beijing.

The strongest version of this narrative is that AI's impact on labor markets is being misunderstood. Rather than a dramatic, job-stealing apocalypse, the real danger lies in the quiet erosion of wages and workplace standards through algorithmic collusion—a subtler but potentially more insidious threat. Ernst’s argument is grounded in empirical evidence, such as the Anthropic study, which highlights the gap between AI’s theoretical capabilities and practical implementation. This perspective usefully shifts the conversation from sensationalist fears to systemic risks, acknowledging that AI’s effects are sector-specific and constrained by real-world factors.
However, the framing of "algorithmic collusion" as the primary threat warrants scrutiny. While the term suggests a coordinated, intentional suppression of wages, the mechanism remains underspecified. Is this collusion emergent from market dynamics, or is it being actively engineered by employers? The narrative also risks downplaying AI’s long-term disruptive potential by focusing on current adoption barriers, which may change as technology and regulations evolve. Additionally, the claim that youth unemployment is "mostly related to the current economic slowdown" could be seen as a false binary—economic slowdowns and technological shifts often interact in complex ways.
Root cause: The paradigm here assumes that AI’s labor market effects are primarily economic and technical, rather than political or social. It echoes historical patterns where technological change is initially overhyped, then underestimated in its systemic consequences. The unstated assumption is that markets and regulations will naturally correct for AI’s downsides, which may not account for power asymmetries between employers and workers.
Implications: If algorithmic collusion becomes widespread, it could further concentrate economic power in the hands of employers, eroding worker bargaining power. The second-order consequence is a potential normalization of wage suppression under the guise of "efficiency," with AI serving as a tool for rent-seeking rather than productivity growth. Human agency is at risk if workers lack transparency into how AI-driven decisions affect their livelihoods.
Bridge questions: How might algorithmic collusion manifest in practice, and what evidence would confirm or refute its existence? What policies could mitigate the risks of AI-driven wage suppression without stifling innovation? If AI’s adoption barriers are overcome, how might its labor market effects differ from current projections?
Counterstrike scan: A coordinated influence campaign pushing this narrative might aim to downplay AI’s disruptive potential to avoid regulatory scrutiny or labor organizing. The actual content, however, presents a nuanced argument backed by evidence, without signs of manipulation. It does not match the pattern of a bad-faith attack.
Patterns detected: none