Former US President Barack Obama recently weighed in on the artificial intelligence debate, asserting that the current buzz surrounding AI is far from “overhyped”. While the financial valuations and commercial promises of AI companies might be exaggerated, Obama argues that the core technology and its profound implications are entirely real.

An Ignored Early Warning

Obama noted that he has been closely tracking AI’s trajectory for years. During the final years of his presidency in 2015 and 2016, he convened a commission of top experts to produce a comprehensive report exploring not only the technological aspects of AI but also its political, social, economic, and ethical implications. Unfortunately, this proactive framework was essentially shelved and ignored by the administration that followed, much like his early pandemic proposals.

The Era of “Recursive Learning”

What makes AI so profound today is its staggering rate of acceleration. Obama explained that we have now entered the phase of “recursive learning,” where machines no longer have to wait for humans to figure things out; they are actively teaching themselves.

He illustrated this shift with a stark comparison: just a year ago, roughly 90% of an AI model’s learning was guided by human “teachers,” with the machine filling in the remaining gaps. In less than six months, that ratio has shifted to a 50/50 split. If current trend lines continue, we will soon see models teaching themselves 90% of the time. Humanity is currently hitting a “hockey stick trajectory,” where AI systems are getting smarter at an exponential pace.

The True Threat: Misalignment, Not Sci-Fi Tropes

As AI models become exponentially smarter, many fear the classic science fiction scenario: superintelligent killer robots deciding humanity is unnecessary. While Obama acknowledges there is a “non-zero chance” of such an apocalyptic event, it is not the risk that concerns him the most.

The primary danger is not that computer models will suddenly develop consciousness or a deep malice toward humanity. Instead, the real threat lies in what experts call “misalignment”. As these rapidly advancing systems begin to set their own agendas, a dangerous gap can form between what the AI wants to do and what humanity actually needs it to do. Managing this misalignment will be the critical challenge as we navigate the accelerated future of AI.