The AI race is in its full swing, and Sundar Pichai is aware of the possible outcomes if things go wrong. In a recent interview with Matthew Berman on the Forward Future platform, Google and Alphabet CEO Sundar Pichai spoke about a fundamental issue with modern technology, stating that the single biggest risk currently facing AI development is the unyielding “race condition” between frontier AI labs.
During the conversation, Pichai highlighted that the constant pressure to compete and leapfrog rivals has locked these tech giants and frontier research institutions into an irreversible sprint. “The real danger starts when slowing down is no longer an option,” Pichai observed, pointing to a critical inflexion point where the sheer momentum of innovation risks outpacing human oversight.
The AGI ‘race condition’
The term “race condition”, traditionally used in computer science to describe a flaw where multiple processes concurrently execute without proper synchronisation – resulting in unpredictable behaviour – is now used by Pichai as a metaphor for the current state of AI development.
With leading AI labs like Google, OpenAI, Anthropic, and Meta constantly pushing out major model upgrades, the window to thoroughly evaluate safety protocols, alignment risks, and downstream societal impacts is shrinking drastically. There’s less time for testing as the labs and firms are not willing to release the latest model before one of the competitors takes the step..
Pichai noted during the interview that this relentless cycle distorts the environment for responsible deployment. “Every few weeks, the perception of who’s ahead changes,” he explained. This shift in public and investor sentiment forces labs to prioritise rapid, frequent releases to maintain market dominance rather than taking a step back to carefully evaluate the broader consequences of their systems. Consider the Claude Mythos as an example – Anthropic had to consider a staggered release to keep it from going into unsafe hands.
Competition becomes a societal risk
The true threat, however, lies in the widening gap between technological advancement and institutional governance. “Once AI systems start accelerating faster than institutions can respond, competition stops being a market dynamic and becomes a societal risk,” Pichai warned.
When regulatory bodies, educational frameworks, and legal systems cannot adapt quickly enough to the shifting baseline of AI capability, society is left exposed to unchecked structural vulnerabilities. Various issues ranging from sophisticated AI-driven cyberattacks to automated economic disruption are no longer just theoretical cases – they are capable of putting immediate systemic pressures, which, when amplified by market demand for speed over safety, make it more dangerous.