It was a fine spectacle to watch the Nevada Legislature try to hogtie artificial intelligence like some wild steer fresh out of the mountains when AI is more like fog–and slips through your fingers and makes you look foolish for trying.
During this session, more than a dozen bills got cooked up, all promising to regulate, restrict, survey, or otherwise boss around the machines now clever enough to answer questions, settle arguments, and, soon enough, probably run for office themselves. The grand hope seems that if enough laws get passed quickly enough, maybe the government can pretend it was steering the wagon the whole time–instead of hanging on for dear life.
The Department of Employment, Training, and Rehabilitation, being the first to admit it’s got more work than workers, put an AI system from Google to work, ruling on unemployment appeals, with a human standing by to nod solemnly at whatever the machine decided. Over at the DMV, an AI chatbot got set loose to answer questions, and from all accounts, it has performed no worse than the flesh-and-blood clerks it replaced–which is to say, it left most folks equally confused.
Not wishing to be left behind, the Office of the Chief Information Officer announced a “responsible and ethical” AI policy last fall–a document so carefully worded–that it kept a bureaucrat busy enough to miss lunch. They banned the use of AI for discriminatory content, demanding that all personal data used by the machines get washed clean first–a fine idea, assuming anyone can figure out what “clean” means when it comes to a machine that knows your shoe size, your dog’s name, and what you ate for supper last Tuesday.
The same office declared that every state agency would soon get its own Microsoft-powered AI assistant–good for writing emails and “brainstorming,” which in government circles means having a meeting that accomplishes nothing but burns up the afternoon.
Not to be outdone, U.S. Senator Dina Neal put forth SB199, a bill as broad as a river in flood season. It started life banning AI-written police reports and AI-built lesson plans for schools, and following some horse-trading and watering-down, it now mostly orders the creation of “working groups” to study the matter–a sure sign that the problem has been declared too complicated to solve and will thus be talked to death instead.
The bill does manage to protect folks’ private medical data from being fed to insurance company machines, and it forbids landlords from using secret AI tricks to boost your rent behind your back–a practice already alive and well across the California border, where AI has been whispering sweet nothings into landlords’ greedy ears.
Meanwhile, Nevada’s Division of Welfare and Supportive Services is fixing to use an AI-powered grocery app that tracks wasted food and serves up deals to SNAP recipients. Thus, while AI won’t plan your child’s lesson in history class, it might soon be trusted to tell you where to find half-priced ham.
Other bills are riding along–one to ban AI from denying your medical care, another to keep AI from making final decisions in emergency management, and another still to outlaw AI-generated child pornography–a sickening thing to have to write a law about, but necessary when humans and machines conspire.
There was even a bill to help voters figure out when a political ad is gettin’ spun by a machine, which is just as well, seeing as some human campaign ads are already so low-down and dishonest that a robot could hardly do worse.
A few bills died quiet deaths before the first committee deadline, but not to worry–their language and good intentions got grafted onto surviving bills, like mismatched planks hammered onto a boat that’s already halfway sunk.
Nevada’s attempt to rein in AI looks much like trying to put trousers on an octopus–messy and guaranteed to leave everyone inked and wondering whose bright idea it was in the first place.
The Old Blog in the Desert 
Leave a comment