Most experts agree that current AI systems aren’t quite Artificial General Intelligence, but mathematician Terrance Tao has an interesting term for what they are: Artificial General Cleverness.
The Fields Medal winner and UCLA professor shared his perspective on the current state of AI in a recent post, offering a nuanced view that challenges both the hype and skepticism surrounding modern AI systems. His observation comes from direct experience—Tao has been actively incorporating AI tools into his mathematical research, including work on notoriously difficult problems from the Erdős collection.
Cleverness Without Intelligence
In his post, Tao draws a careful distinction between what AI systems can actually do today versus the aspirational goal of artificial general intelligence. “I doubt that anything resembling genuine ‘artificial general intelligence’ is within reach of current #AI tools. However, I think a weaker, but still quite valuable, type of ‘artificial general cleverness’ is becoming a reality in various ways,” he says.
Tao then expands on what he means by general cleverness. “By ‘general cleverness’, I mean the ability to solve broad classes of complex problems via somewhat ad hoc means. These means may be stochastic or the result of brute force computation; they may be ungrounded or fallible; and they may be either uninterpretable, or traceable back to similar tricks found in an AI’s training data. So they would not qualify as the result of any true ‘intelligence’. And yet, they can have a non-trivial success rate at achieving an increasingly wide spectrum of tasks, particularly when coupled with stringent verification procedures to filter out incorrect or unpromising approaches, at scales beyond what individual humans could achieve,” Tao continues.
“This results in the somewhat unintuitive combination of a technology that can be very useful and impressive, while simultaneously being fundamentally unsatisfying and disappointing – somewhat akin to how one’s awe at an amazingly clever magic trick can dissipate (or transform to technical respect) once one learns how the trick was performed,” Tao continues.
“But perhaps this can be resolved by the realization that while cleverness and intelligence are somewhat correlated traits for humans, they are much more decoupled for AI tools (which are often optimized for cleverness), and viewing the current generation of such tools primarily as a stochastic generator of sometimes clever – and often useful – thoughts and outputs may be a more productive perspective when trying to use them to solve difficult problems,” Tao says.
AI as a Mathematical Research Tool
Tao’s perspective isn’t merely theoretical—he’s been putting AI to work in his own mathematical research with notable results. The mathematician has been using AI systems to tackle problems from Paul Erdős’s famous collection of mathematical conjectures, some of which have remained unsolved for decades.
Beyond problem-solving, Tao has found AI valuable for evaluating potential solutions and exploring different approaches to proofs. This practical application aligns perfectly with his concept of “artificial general cleverness”—the AI systems generate multiple approaches through what he describes as “somewhat ad hoc means,” which Tao can then verify and refine using his mathematical expertise.
The Magic Trick Analogy
One of Tao’s most striking observations is his comparison to magic tricks. Current AI systems can produce impressive results that seem almost magical, but understanding how they work—through pattern matching, statistical associations, and brute force computation—can transform that awe into something more measured. This doesn’t necessarily diminish their utility, but it does ground expectations about what these systems are actually doing.
The key insight is that for AI tools, cleverness and intelligence may be fundamentally different things. While humans tend to associate the two traits closely, AI systems optimized for cleverness don’t necessarily exhibit what we’d recognize as true intelligence or understanding.
A Framework for Using AI
Tao’s framework suggests a practical approach: view current AI systems as “stochastic generators of sometimes clever—and often useful—thoughts and outputs.” This perspective acknowledges both their capabilities and limitations. It suggests that the most effective use of these tools involves generating multiple potential solutions or approaches, then applying rigorous verification and human judgment to separate the promising from the flawed.
For businesses and researchers grappling with how to integrate AI into their workflows, Tao’s perspective offers a grounded middle path. These systems aren’t the artificial general intelligence of science fiction, but they’re also more than simple automation. They’re tools that can augment human problem-solving by exploring solution spaces at scales humans couldn’t match alone—as long as human expertise remains in the loop to verify and validate the results.
As one of the world’s leading mathematicians actively using these tools in his work, Tao’s assessment carries particular weight. His concept of “artificial general cleverness” may provide a more accurate and useful framework for understanding—and leveraging—the current generation of AI systems than either the hype of imminent AGI or dismissive skepticism.