The Transportation Department, which oversees the safety of airplanes, cars and pipelines, plans to use Google Gemini to draft new regulations. “We don’t need the perfect rule,” said DOT’s top lawyer. “We want good enough.”
This looks like a notoriously bad idea.
In 2023, a city (Porto Alegre) near-ish my homeland approved a law initially proposed by ChatGPT, then manually reviewed and edited. Here’s a link; it shows both the initial proposal and final version (both in Portuguese).
The law addresses some shite the water and sewage department (DMAE) did often:
- install new water meter for a house, with no regards to its placement or securing it properly
- wait until water meter gets stolen for parts (welcome to Latin America!)
- charge house owner for a new water meter
- go back to step 1.
So a councilperson prompted ChatGPT to draft a law addressing it. And the draft sounds reasonable… until you inspect it further, and notice a certain article omitted from the final revision:
[Rough translation] 7th article. DMAE shall be allowed to establish complementary norms to regulate the enforcement of this law.
Why was this article omitted? Remember: DMAE was the very department being legislated against. If allowed to issue “complementary norms” regarding that law, the law would become toilet paper — because all the department had to do is to claim “the law is only valid if the theft happens in the 31st of February!” or some equally dumb shit.
The issue I mentioned above was fairly specific, the solution was straightforward, and mostly non-partisan. And the entity in question was a city government, so no “nested” political entities. And the e-muppet was still able to drop such a huge bollock.
What would happen if this was done on a country level? And it included partisan matters? And the issue was something complex, with no “right” answer?
That’s what I’m thinking, while reading the link in the OP.
So we’re asking ChatGPT to write laws now?
You can tell this isn’t true because there is no way trump plans to do anything with any kind of intelligence.
best move for deregulation
Zerzan appeared interested mainly in the quantity of regulations that AI could produce, not their quality. “We don’t need the perfect rule on XYZ. We don’t even need a very good rule on XYZ,” he said, according to the meeting notes. “We want good enough.” Zerzan added, “We’re flooding the zone.”
They are going for more regulation. If I had to guess, probably to flood federal courts with cases, hoping that they are too overwhelmed to stop anything important.
It’s anticompetitive. Trying to keep up will kill off many businesses, which can be snapped up by the larger firms and corpos that can either afford to pay the fines or can simply ignore new regulations while their lawyers try to fight.
Oh my god, this isn’t The Onion.
