Looking at actual bill, it proposes:
- Creating a paid "no tracking" option. Obvious failure mode: the option price is $100M, and anybody who uses it is unable to use 99% of the functionality, since it requires some form of tracking. Obvious next step - the law requiring this option to be no more than 10% more expensive than regular membership. Obvious next failure mode: inapplicable to sites that do not charge for membership. Obvious next step - creating a government commission empowered to decide what sites are supposed to charge for "no tracking" membership and which services it is supposed to cover. If you like your site subscription - you can keep your site subscription.
- Penalize large companies that submit false information in their annual privacy report - thought submitting false information to any government agency is a crime anyway? And for a public company, I assume publishing almost any false report would immediately put them under the shadow of fraud charges from SEC. So declaring something that is already a crime a crime again is supposed to... what?
- Require companies to assess their algorithms for accuracy, fairness, bias and discrimination. Obvious failure mode: who does the assessment? Obvious next step: creating a government commission empowered to approve algorithm fairness assessment standards. Obvious failure mode: since nobody knows what "fairness" it, it turns into another partisan tug of war, to be used as a club against companies affiliated with opposite tribe or just representing a good jumpstart to the next political campaign. Reasonable academic discussion of algorithmic bias becomes impossible, buried under layers of partisan tribal rhetoric and professional offense miners. Billions are spent annually on "bias prevention", without any shade of solution on the horizon, on the contrary, the problem becomes worse every day - at least if you're listen to bias prevention industry, but they're the only ones who are allowed to speak on the topic.