EdBoard: Shapiro's Executive Order on AI is a good start
On Wednesday, Sept. 20, Pennsylvania Governor Josh Shapiro visited Carnegie Mellon to sign an executive order that establishes guidelines for the use of generative AI in the state government. Shapiro spoke with President Jahanian, along with other representatives from his administration, on the immense potential of generative AI.
Shapiro claims it is the "most comprehensive executive action" that any state government has taken on the matter, citing his desire to "embrace AI, not fear it" and proactively set forth guidelines for its use.
It's refreshing to see a government respond to advances in tech. Public policy often seems to be years behind the curve, but this doesn't need to be the case; we should expect our governments to react to breakthrough, paradigm-shifting technologies in a thoughtful, progressive way.
The Executive Order has three primary mandates:
Establish a "Generative AI Governing Board": This is a 12-person body composed of administration officials and experts in the field who will "develop thoughtful policies" on how generative AI may be used and incorporated into the daily work of government employees.
Define core values: These are values the government agencies ought to adhere to in their use of generative AI, including privacy, accuracy, equity and fairness, and employee empowerment.
Create a framework for training: The administration is developing programs to train and certify government employees on the use of generative AI. In addition, they will establish a 2-year fellowship program to "recruit more experts in this field."
While the order itself is rather limited in scope, it's a good model for the attitude with which we should approach AI. The training and certification program will only apply to commonwealth employees, and the "core values" it establishes are little more than guidelines, but these are good places to start. However, we should expect this preliminary set of policies to be followed by more comprehensive legislation. While a tool with immense potential, generative AI still raises many thorny questions in the realm of privacy, copyright, bias, and ethics. The order also doesn't establish guidelines on AI use in education. Most national governments have no policies on the use of AI in education, instead giving individual schools the discretion to choose their own policies. However, most schools and universities have formal guidance on the matter either, many of which (including Carnegie Mellon) give individual professors the discretion to choose how to use it in their courses.
We should approach AI with an attitude of skeptical optimism. To ignore (or even outright ban) generative AI may leave the technology in a legal limbo, making us all more vulnerable to its misuse. It's good to see that the government is reacting to this technology and discussing its use in a nuanced way, but we should keep in mind there are still many issues that will have to be addressed as it evolves and becomes more integrated into our lives.