Executives at the world's leading artificial intelligence laboratories are making an increasingly bold claim: their most advanced models have moved beyond serving as mere tools and are now actively contributing to the development of next-generation systems.
Sam Altman, CEO of OpenAI, recently stated on the "Relentless" podcast that "we are already, roughly, inside the singularity." Demis Hassabis, founder of Google DeepMind, described the field as standing "at the foothills of the singularity," a term referring to a theoretical point where technological progress accelerates at an unpredictable and potentially uncontrollable pace.
Anthropic similarly claims its Claude models already assist in developing more powerful AI systems. According to this perspective, models' abilities to write code, analyze experimental results, identify vulnerabilities, and solve scientific problems could create an acceleration loop: each generation helps engineers and researchers build the next one faster.
Yet the reality may be more modest than the rhetoric. Business Standard explains that while AI adoption is indeed accelerating, all of that expansion is happening under the watchful and purposeful eyes of humans. Today's AI expansion remains highly dependent on human-controlled capital, chips, electricity, data centers and deployment decisions.
Altman's own more careful framing contradicts the bolder headlines. In his June 2025 essay titled "The Gentle Singularity," he described a "larval version of recursive self-improvement" and was explicit that this "isn't the same thing as an AI system completely autonomously updating its own code." That distinction matters. Assisting engineers in building better systems differs substantially from systems autonomously rewriting their own architecture without human intervention.
Expert consensus, meanwhile, remains far more conservative. A December 2025 analysis by AI Multiple aggregating thousands of forecasts concludes that surveyed AI experts "estimate that AGI will probably emerge between 2040 and 2050 and is highly likely to appear by 2075," while headline predictions from industry executives place singularity around 2026. The gap between insider optimism and expert consensus has widened rather than closed, as reported by Forbes.
If we're already at the singularity, AI advancement should be accelerating rapidly without human oversight or control. Yet while AI use and adoption is indeed accelerating, all of that is happening under the watchful and purposeful eyes of humans. The technology remains a tool, albeit an increasingly powerful one, still operating within human-defined guardrails. Whether that distinction will hold as the systems grow more autonomous remains the open question beneath all the singularity talk.
