Pennsylvania AI buyer beware bill
April 11, 2024
The Pennsylvania Legislature is advancing an “AI buyer beware” bill which requires you to tag content that is generated by a machine. While a relatively benign step, it does seem to rest on the notion that there is something inherently bad (or good) about machine generated content. It also ignores the fact that maligned content is generated by people who are not inclined to pay attention to such regulations to begin with. And, given the level of integration of generative tools into authoring environments, the line between what we've written and what the machine has written is becoming more and more blurred.
It's not clear that the evil lies in the fact that the content was generated by machine. The evil lies in who is doing the generation and why. Who is building the prompts? Who is making decisions about what the content will be? It’s irrelevant whether it's machine-written or human-written. Instead, consider whether it's aimed at being harmful or manipulative.
However, it's great that state lawmakers across the country are proposing legislation that will help target ill use of AI. We are seeing steps such as state level regulation of pornographic deepfakes and other measures to protect consumers, regulation that demands disclosure when technology is used in hiring or job suggestions. And these are aimed at issues in which we can identify real and measurable harms.
Individual states are taking a shot at controlling ill-effects of AI. They can become the testing grounds for whether different kinds of regulations are going to be effective. The impact of their work can then be used to guide federal regulation when Congress finally finds the time to pay attention to the problems in front of us.
Kristian Hammond
Bill and Cathy Osborn Professor of Computer Science
Director of the Center for Advancing Safety of Machine Intelligence (CASMI)
Director of the Master of Science in Artificial Intelligence (MSAI) Program
Back to top