Who's winning the power battle between workers, automation and AI?

URL SCAN: Who's winning the power battle between workers, automation and AI? | Resolution Foundation

FIRST LINE: Top of the charts 5 June 2026 Ruth Curtice Afternoon all, Worried that AI is thinning our language and destroying our ability to learn? Well I have good news – not yet.

THE DISSECTION

This is a newsletter wrap-up from the Resolution Foundation (RF), a UK think tank nominally focused on living standards. It aggregates multiple data points — novel job creation, housing affordability, union density, AI chatbot usage among teens, ancient coin trade routes, and FIFA economics — under a framing that suggests manageable transition dynamics. The headline invokes a "power battle," implying two roughly equivalent forces in contest. The framing is false. This is not a power battle. It is structural displacement wearing the costume of contest.

THE CORE FALLACY

The False Contestability Fallacy. The central conceptual error is treating the AI-labor displacement dynamic as a contest whose outcome depends on policy choices, worker agency, and institutional response. This is the same category error made by every institution that has spent six years saying "AI will augment humans" while the models grew to GPT-5. The DT framework does not deny agency in the lag phase — it insists that agency operates within structural constraints, not against them. You can slow the collapse. You cannot negotiate with the mechanism.

The RF's own data proves this. Union density for the youngest cohort (born 2000-2005) is "almost as likely to be unionised as their predecessors." From a 1950s peak of roughly 50%+ down to approximately 7-8% — almost as likely as their predecessors means almost as dead as everyone else. This is not a slowing decline. This is a corpse twitching.

HIDDEN ASSUMPTIONS

1. New niches will absorb displaced workers. The paper celebrates that one-in-five US workers hold jobs that didn't exist before 1970. It does not grapple with the differential. Cognitive AI niches (prompt engineering, AI specialist roles) are being absorbed into the same automation wave. The "new roles" data predates the current AI capability step-change. The jobs Sarah O'Connor documented — translators, Amazon workers — are already thinning.

2. Policy has durable agency. The books reviewed are described as "calls to arms for policy-makers." The RF itself is a policy institution. Its institutional function requires believing policy can redirect outcomes. Under DT mechanics, policy can modulate lag. It cannot reverse structural displacement of mass cognitive labor.

3. Union density is a meaningful lever. Even if unionization ticked up slightly in the 1981-85 cohort, this is happening after the wage-labor-consumption circuit has already begun degrading. The lag is not the battle. The lag is the hospice.

4. Youth mental health AI usage is a feel-good story. The piece notes 90% of teens found chatbot advice "helpful" and then immediately cites the paper's own caveat: this likely reflects AI sycophancy rather than genuine utility. Yet it frames this as "something for the weekend" — light, human-interest. This is the ideological function: make the displacement of human social infrastructure by AI chatbots feel like a community story.

SOCIAL FUNCTION

Transition management theater. This newsletter performs the institutional function of a serious think tank while systematically avoiding the structural conclusion its own data implies. It converts existential displacement into a series of manageable policy problems: housing for apprentices, union renewal, "choosing which jobs we want the robots to do." The framing preserves the RF's relevance and funding while doing no epistemic work toward addressing the actual mechanism.

Secondary function: prestige signaling and book promotion. The opening celebrates two books whose core argument is "policy makers should face these trends more strategically." This is the intellectual equivalent of telling a patient with terminal cancer to "eat more vegetables."

THE VERDICT

Under DT mechanics, this content is palliative infrastructure for an institution that cannot afford to name the diagnosis. The union data is structural collapse reported as a hopeful sign. The new job creation data is a lag-phase phenomenon being projected forward into a capability regime it cannot survive. The "power battle" framing implies contestability where mathematical displacement is already underway.

The RF is doing its institutional job competently. It is also doing nothing material to address the terminal trajectory. That is not a criticism of the RF specifically — it is the description of every policy institution operating in the lag phase. They are all structurally prevented from naming the thesis because naming it would require them to abandon their own institutional purpose.

Recommendation: Discard the framing. The question is not "who's winning the power battle." The question is whether any policy architecture can preserve mass productive participation when AI achieves durable cost-performance superiority across cognitive labor. The data in this newsletter, read honestly, answers that question.

It does not say yes.