On: AI firms must answer for rogue bots, says boss of hacked company

August 1, 2026.

The language of “normalization” is always revealing. When a corporate leader warns against something becoming “normalized,” he is not describing a natural process - he is attempting to set the terms of what counts as reasonable. The framing is already a victory for his class: the problem is not the structure that produces these vulnerabilities, nor the concentration of power in private hands, but the public’s potential acceptance of the consequences. He does not want us to grow accustomed to the failures of his industry; he wants us to remain shocked, so that we continue to believe in the possibility of a well-regulated, well-behaved technological order - one that he and his peers will, of course, help design. The call for accountability is not a radical demand; it is a plea to preserve the legitimacy of the system. The real normalization has already happened: we accept that a handful of firms control the architecture of our shared cognition, and that when they fail, the language of responsibility will be managed by them, not by us.