Voice and Exit
Regulators seem to have a tough time anticipating how people will react to their behavior.
Europe Is in Danger of Regulating Its Tech Market Out of Existence
European officials seem shocked that US tech giants would rather take their ball and go home than put up with poorly made rules and greedy regulators.
‘The economist Albert Hirschman once described the two options in an unfavorable environment as “voice” and “exit.” The most common option is voice—attempt to negotiate, repair the situation, and communicate toward better conditions. But the more drastic option is exit—choosing to leave the unfavorable environment entirely. That’s more common for people or political movements, but it’s growing increasingly relevant to technology in Europe.’
U.S. Needs Tech Regulation Unification, Experts
Regulatory incoherence within just the federal government inhibits international collaboration.
‘The United States needs a unified approach to tech regulation when approaching the global stage, said experts at an American Consumer Institute panel discussion on Tuesday.
‘The U.S. faces challenges negotiating international agreements due to the lack of a coherent position within the government, said Simon Lester, founder of WorldTraidlaw.net, an online library that provides resources on global trade. "If you're negotiating an international agreement, it's crucial that you have a clear, unified point of view within your government."’
Preview of a Post-Chevron World: The Seventh Circuit Upholds Regulation B's Discouragement Prohibition as Consistent with ECOA
Apparently, if the statute is clear in its intent, then there is no Loper issue. Congressional delegation of authority to the agencies is fine, subject to certain circumstances. Also, the article suggests that deviating from statutory intent to reflect industry realities may be okay, too.
It’s not exactly cats-and-dogs-sleeping-together chaos, is it?
‘The Seventh Circuit found Regulation B’s discouragement provision consistent with the ECOA statutes and Congress’s statutory directive/prohibitions. Considering ECOA’s text “as a whole,” the court found that it prohibits not only “outright discrimination” against credit applicants, but also the “discouragement of prospective applicants for credit.” The Court noted that in ECOA § 1691b(a), Congress conferred upon the agency broad authority to enact regulations to carry out ECOA’s purposes, including regulations “to prevent circumvention or evasion thereof, or to facilitate or substantiate compliance therewith.” Conferring authority to prevent “circumvention or evasion” indicates that ECOA “must be construed broadly to effectuate its purpose of ending discrimination in credit applications.”’
These landmark Supreme Court rulings could throw AI regulation into turmoil
Regardless of who wins the Oval Office this Fall, their hands may be tied thanks to Loper, Jarkesey, and Corner Post.
‘In Heflin’s own opinion outside of his work at Trustible, he said the power of the executive branch may come close to a complete halt.
‘Two choices exist: to not make progress on rulemaking at all, or to write something so narrow that may not capture all of the possible harms at play. AI presents several high-risk use cases, like being employed in healthcare, education and human services. ‘
Rationing for Sickle-Cell Therapies
Drug company develops life-saving treatment for the thousands of Americans who suffer from sickle-cell disease, an affliction that causes most victims to die before the age of 50. The drug is expensive, costing millions per patient.
CMS proposed paying for the drug on two conditions: the drug company rebates some of the price to the government if the promised results fail to arrive, and the drug company pays for fertility treatments to counter one known side effect of the therapy. The manufacturer agreed to both terms conditionally. They wanted an opinion from the government that other parts of the healthcare bureaucracy wouldn’t charge them with crimes on the books intended to prevent bribery. The fertility treatments might look like a payoff.
The IG said that the fertility program was illegal, but he wouldn’t write an opinion.
What the right hand gives, the left hand takes away.
‘Alas, the HHS Inspector General told Vertex executives in January that its fertility program is likely illegal, though it refused to issue a written opinion. Perhaps because an opinion would reveal how the agency is using dubious legal interpretations to deny life-saving treatments.
‘Vertex last week sued the government, arguing the Inspector General’s position is arbitrary and capricious. The government’s view is that offering fertility preservation services to counter a gene therapy’s side effect will somehow induce sickle-cell patients to get a life-saving treatment that they shouldn’t get. This makes no sense.’
The Billion-Dollar Government Mandate You Probably Haven’t Heard About
Perhaps there will be a new AI app that can deal with the compliance overhead from the prevailing wage laws: requirements for contractors on government work to follow fluid and byzantine compensation rules for their employees. Until then, we have litigation.
‘But, as decades of data show, prevailing wage laws hurt everyone else. They’re notoriously difficult to implement in the field, forcing contractors to painstakingly track and classify employees’ tasks (for example, paying a general laborer as a “carpenter” if he happens to hammer a nail that day). They hurt employees, particularly entry-level ones, by making it punitively expensive and complicated to hire workers. The brunt of it falls disproportionatelyon minorities, immigrants, younger workers, women, veterans, and small businesses. And they cost taxpayers more by excluding qualified businesses from competing for public-works contracts and driving up costs (not only payrolls, but compliance costs) for those that remain.’
Regulations Targeting Large Language Models Warrant Strict Scrutiny Under the First Amendment
If Communist China can direct its Generative AI to express what they see as the right idea, then it’s straightforward to see LLMs as deserving of First Amendment protection. The data that trains these models and the decisions made on their architecture are important things to consider. If this is true, then regulation of AI could be less than industry worries.
‘LLMs, therefore, present a straightforward case for protection as pure speech. If the inputs to an LLM are all expressive activity, and if the editorial judgment that goes into curating those inputs is expressive activity, LLMs should, in short, be no less deserving of First Amendment protection than the sum of their parts.’