Humans Can Be Held Accountable. What About Corporations and AI?

In the course of human events, it is necessary to hold people accountable for their actions and decisions. But as we extend personhood to corporations and cognitive load to AI chatbots, the line between what entity is making what decision becomes unclear. Attorney generals are seeking to criminally implicate corporations for their actions, including those that host large language models, but corporations and AI chatbots cannot serve prison sentences for criminal penance.

So, is our natural sense of accountability developed over millennia wrong, or are our current systems? Or are we forced to live in a perpetual state of cognitive dissonance as with many other issues in our modern life?