This is a really important point, and I think it derives from a fundamental asymmetry. Employees have ~1 job, but employers have lots of employees. The end of an employee-employer relationship is necessarily way harder on the employee than the employer. Employers can cause way more trouble for employees than vice versa. And employers can amortize the cost of trickery.
A good example is contract negotiation. Employers can sneak in terms that many employees won't understand. The ones who do understand have a much harder time fighting, and pay a much higher proportional cost to do so. Hiring a good employment lawyer to help negotiate could easily cost 10% of a median employee's annual salary, and a lawsuit could cost 5-10x a median salary. (For lower-paid workers, it's proportionally even worse.) But those same costs for employers are generally negligible.
Market outcomes are much more likely to be optimal when you have many agents of approximately equal power. As you move toward strong power asymmetries (that is, toward monopolies and monopsonies) you get much lower odds of optimal outcomes.
So the common counter to this is where employees band together to negotiate as a unit: unions. And regulations let us rule out things that it's basically a waste of time for everybody to negotiate, or where allowing negotiation will lead to systemically poor outcomes. For example, CA basically forbids noncompete clauses in worker contracts. This is possibly bad for individual employers, but it's way better for employees and employers as a whole, because it increases innovation and therefore total wealth.