Raw LLM Responses

I could see it as being more of an attribution model, where each "robot" needs to be assigned a value, as part of the the company's overall valuation. Then from there, you could assign the equivalent of pension contributions to each robot / software entity, which the company would obviously pay on their "behalf". (Remember that employee payroll deductions are where most taxes get paid - it's not corporate taxes). And a robot might only be held liable for its own actions if the company "claimed" it and did source deductions against the comparable value each year. Otherwise the liability just lands on the company. I'm not saying this is a sure thing, it just seems like a natural course of action for money-hungry governments facing a shrinking (human) workforce.