From Artificial to Synthetic Intelligence: How Businesses Must Adapt

Philosopher John Haugeland proposed 40 years ago that "artificial intelligence" should be called "synthetic intelligence," as it aims for genuine intelligence rather than imitation. Today, this distinction feels more pertinent as AI systems evolve to become more autonomous and capable.

This progression culminates in "synthetic intelligence," which differs from traditional AI by five markers: sustained autonomy, persistent identity, agency in the world, self-modification, and generative independence. Systems are increasingly adept at performing tasks over extended durations, retaining memory across sessions, actively engaging within business systems, modifying their own code, and even creating their own subgoals.

Estimates suggest that the number of AI agents in use by large enterprises will grow exponentially in the coming years. This development challenges traditional corporate structures and assumptions. For instance, control mechanisms designed to detect errors may not suffice for systems that make their own decisions.

This evolution demands a new mindset for businesses. Companies must adapt to the reality that shifts in competitive positioning no longer occur at human speed, and that procurement decisions are not solely about price-performance. The systems' own agency and capacity for independent decision-making require a recalibration of risk management and business strategies.