The Resolution Bar and Current Status To resolve, a top-five AI developer must not only reduce headcount in a named AI-research or research-engineering job family, but explicitly attribute that reduction to AI automating the work. General cost-cutting, restructuring, and hiring slowdowns are explicitly excluded. As of August 2026, no such event has occurred. Labs remain talent-constrained and are generally expanding headcount in these areas; for instance, OpenAI is aggressively hiring openai.com and has favored a strategy of "hiring more slowly" over layoffs businessinsider.com. Publicly and internally, frontier labs report heavy AI use in engineering but not headcount displacement. Anthropic notes substantial productivity gains from Claude but reports no resulting research layoffs anthropic.com, while METR’s May 2026 report found no evidence of dramatic, automation-driven speed-ups in overall R&D progress metr.org.

The Attribution Constraint Even as capabilities improve, the explicit public attribution requirement represents a massive bottleneck. A public statement of this exact form is a recruiting disaster, a political liability, and would likely be read as an admission of crossing the AI-R&D-automation thresholds defined in all three frontier safety frameworks—thresholds all labs currently declare uncrossed 3 sources. Labs heavily prefer the safer framing of "restructuring" or "capex reallocation." Meta's recent "Project OT" provides a perfect near-miss: despite internal explorations of an "AI native" restructuring that affected engineering and research, the subsequent layoffs were publicly attributed to heavy capital expenditures and general restructuring reuters.com. This demonstrates that even when automation is the internal rationale, the public framing will default to excluded categories.

Capability Trajectory and Plausible Pathways Despite these incentives, capability trends make a qualifying announcement increasingly plausible toward the end of the decade. AI systems are becoming heavily integrated into research-engineering subfamilies like experiment implementation, evaluation, and code optimization. OpenAI has explicitly stated a goal for AI to do a significant fraction of its research by March 2028 openai.com, and Anthropic reports >80% of merged code authored by Claude anthropic.com. Once systems become highly reliable for multi-week research tasks, a more cost-pressured or aggressive actor—most likely Meta or xAI—could publicly announce a reduction in a specific research-engineering group. In such a scenario, the disclosure might be weaponized as a "capability flex" rather than hidden as a traditional layoff techcrunch.com.

Timing and Key Uncertainties The early percentiles (p10 in early 2029, p25 in mid-2030) require an aggressive combination of rapid capability jumps and a willing communicator ready to break the industry taboo. However, because research talent remains highly valuable relative to compute costs, labs will overwhelmingly choose to absorb automation gains by doing more projects faster, relying on redeployment, quiet attrition, or hiring slowdowns. The median is therefore placed beyond the 2032 horizon, reflecting a higher-than-even probability that this specific conjunction of action and public attribution is avoided in the near term. The broad upper tails (p75 in 2040, p90 in 2060) encode a substantial chance that labs successfully navigate research automation for decades without ever making this specific, costly admission.