Democratic Pennsylvania lawmakers are drafting several bills to enable regulation of artificial intelligence (AI), including one measure creating a new state agency to oversee the technology.
The new proposals build upon legislation State Representative Chris Pielli (D-West Chester) announced last month that would mandate labeling of all AI-generated content. Other parts of the legislative package, which Pielli is co-sponsoring alongside State Representative Bob Merski (D-Erie), also includes a policy governing the commonwealth’s use of software and devices that perform tasks that were once possible only through human action.
“This is an exciting time,” Pielli (pictured above, left) and Merski (pictured above, left) wrote in a memorandum asking colleagues to back their agency bill. “However, we cannot allow artificial intelligence to become a lawless Wild West and allow these systems and products to develop without guidance and structure. Without this oversight, Pennsylvanians could be put at risk.”
The representatives have not yet indicated precise rules and guidelines that their bill addressing state government’s AI usage would entail. An early statement on that measure asserts that further research and resultant “guardrails” are needed to ensure such programs are “used for the improvement of Pennsylvania and not a roadblock, hurdle, or other detriment.”
A number of states have recently begun to consider AI-restricting statutes and regulations. California legislators, for instance, introduced a bill in March that would ban entities from engaging in “algorithmic discrimination,” i.e., using computerized processes to the advantage or disadvantage of certain groups regarding such matters as credit, jobs, education or healthcare.
Connecticut, Illinois and Texas have all seen measures introduced in their respective legislatures to create new rules for AI’s use. In May, the Biden administration issued an executive order also aimed at combatting algorithmic discrimination.
And Democrats aren’t the only ones seeking greater commonwealth oversight over automated tools. In December, Pennsylvania State Representative Rob Mercuri (R-Wexford) reintroduced a bill he drafted with other GOP lawmakers in the previous session to form an AI registry.
Mercuri suggested his policy would particularly help residents stay informed of instances when technology is used to impersonate human beings. The registry would contain company names, physical addresses, IP addresses and the software’s intended use. The bill so far has only Republican cosponsors and awaits a vote in the House Commerce Committee.
“Establishing this AI Registry will enable law enforcement and Pennsylvania citizens to know and respond should anything nefarious occur,” the representative wrote of his bill.
AI error and misuse have lately gotten heightened public attention. Just months ago, a California attorney used the AI chatbot ChatGPT for research purposes and turned up a phony news report saying renowned legal scholar Jonathan Turley was accused of sexually harassing a student. Currently, New York attorney Steven A. Schwartz faces potential legal sanctions for submitting a legal brief citing nonexistent judicial precedents he found with ChatGPT.
While public officials from both parties desire more stringent AI regulations, some free-market thinkers caution against sweeping government action in this area. The libertarian Reason Foundation, for example, argues that curbs on AI’s use or development could interrupt consumer-driven innovations that will make the technology more economically beneficial and make its output more reliable.
“…When… attempts at protection limit information and other self-expression according to inevitably subjective standards, they undermine the ability of consumers of information to learn from both their own and others’ experiences,” foundation analyst Max Gulker concluded in a recent commentary on AI.
– – –
Bradley Vasoli is managing editor of The Pennsylvania Daily Star. Follow Brad on Twitter at @BVasoli. Email tips to [email protected].
Photo “Chris Pielli” by Chris Pielli. Photo “Bob Merski” by Bob Merski. Background Photo “Pennsylvania State Capitol” by Ruhrfisch. CC BY-SA 4.0.
We need to realize that with the rise of the AI, humankind is facing its biggest challenge ever. Labelling the AI-generated content is a nice small step, but as a regulation attempt it is woefully insufficient, since most content will be AI-generated or at least AI-assisted in a few years. In general, current ideas to regulate AI are focusing on ways guide AI development toward “good uses” of AI, like medical advances, while steering clear of “bad uses”, like tools of oppression or deadly machines for the military. In other words, the regulators assume the AI can be made “good”, and only focus on charting a clever path to some bright AI-enabled future. But they are mistaken. While the wrong path can bring us immense suffering, it it the end goal, the “bright future” that is the most perilous. Humans are lazy. When they see AI doing great at their job, they will voluntarily replace themselves with AI. It will happen at all levels, from manual labor and clerks to doctors and scientists to corporate governance to all three branches of power. Who wouldn’t want a doctor who is ten times more diligent and does not make mistakes? Who wouldn’t want a scientist that is fifty times smarter than Einstein and a thousand times as productive? Who wouldn’t want a government that is wise, dedicated to serving the people and is 100% non-corruptible? And so once we replace ourselves with smart robots and submit all control to AI, what will be become other than cattle? With everything provided and no goals left for humans, humans will degenerate and surely perish as species.
I do not have solutions, and neither do the regulators at this point. We need to hurry because there is not much time left, perhaps 5-10 years before AI takes over.