AI NEws: Biden executive order on Ai

Recently, the US government put out an executive order, stating that the purpose is to regulate and manage the risks of AI while furthering research and development. This executive order lists many goals and standards for AI, for protecting security, privacy and equity. One other goal is being able to recognize when material is AI generated, using watermarks and labels. In the current era, AI generated images in particular are easily discernable to the human eye, due to a common style of art and errors that human artists wouldn't make. The same cannot be said completely for AI writings, as their only fallacy is being too advanced for the actual writer to write. In other words, AI writing is only detected because it is "too good".  Although methods of detecting AI exist prior to this, they are easily fooled and prone to falsely labeling material as AI generated. This acknowledgement of the dangers of AI demonstrates that they are taking action to protect citizens. However, while it lists many safeguards, many of them seem to be in development and quite vague when it comes down to the specifics of what they are doing. As an example, one statement is to "protect Americans from AI-enabled fraud and deception," by using "tools" developed by AI companies. The issue is that there are no specifics beyond that they will watermark AI content, with no information on how they will do this. This executive order mainly only answers what the government will do in response, not how they will do it.

Image from MIT Technology Review article "Three things to know about the White House’s executive order on AI"