GenAI Will Change the World. But There Are Still Things It Can't Do

Generative AI is advancing at an unprecedented pace, and the excitement is justified. When consultants used AI tools, one study showed quality improvements of up to 42%, a figure that rivals the long-term productivity gains from steam and electricity. Yet the same research revealed a striking blind spot: GenAI performed worse than humans at solving specific business problems with right and wrong answers, and training people in AI tools actually made their outputs worse, not better.

This Forbes piece cuts through the hype to examine three underappreciated constraints on GenAI's impact: its dependence on modular business structures, its tendency to homogenise output at the expense of strategic differentiation, and its limitations when applied to problems requiring judgment rather than recombination.

It also offers a practical three-level framework for how organisations should think about integrating GenAI, from raising the productivity baseline, to rethinking the business model, to reimagining the value-add of an entire sector.

Read here.