The only downside is that arbitration is still too much like litigation.
I've been dreaming about a ballot proposition which says that if a person agreed to a contract that enforced binding arbitration, and was never given an opportunity to amend the contract, then the consumer would be eligible to pick the arbitrator from the set of licensed arbitrators. The company still gets its cost savings from avoiding court, but is not in effect buying a whole firm of arbitrators. The consumer at least can get some attempt at a fair choice in a case of a complaint, which is more choice than they have now.
The Supreme Court decided in AT&T Mobility LLC v. Concepcion that the Federal Arbitration Act preempts state laws disfavoring arbitration. Your proposed law would invalidate a large number of existing arbitration agreements and so would likely also be preempted. At this point, it will take an act of Congress to make any significant changes to existing consumer arbitration laws because conflict preemption will invalidate nearly any conceivable state law favoring consumers.
It's an upsetting state of affairs because the federal government does little to protect consumers yet consistently interferes with the ability of state governments to do so (see, e.g., the Office of Comptroller of Currency's frequent interventions to prevent state regulators from investigating predatory lending practices in the years leading up to the financial crisis).
I'm gonna call [citation needed] on this. The recent NYT series on binding arbitration claimed exactly the opposite: that arbitration almost always rules in favor of the company and that arbitration companies which do rule in favor of individuals quickly find themselves getting no more business. They cited peer-reviewed sources and law professors, so I want to see yours.
Would you mind highlighting the article with peer-reviewed sources and law professors explaining that "arbitration almost always rules in favor of the company and that arbitration companies which do rule in favor of individuals quickly find themselves getting no more business"?
Furthermore, even if it was true that "arbitration almost always rules in favor of the company", why would that alone be the sign of a problem? We shouldn't just assume that, for example, half of the victories will go to the company and half to the individual; we shouldn't assume any particular percentage breakdown without more information about who initiates arbitration and why. (The "priors", to make an analogy to Bayesian statistics)
Rejecting the offered source with a "tl;dr" complaint is an invalid distraction from that request.
Not to mention, many topics have not been the subject of peer-reviewed research; if you require peer-reviewed research as a necessary condition on you being persuaded about anything, then you are severely limiting the scope of potential debate.