On Oct 30, 2023, President Joe Biden issued the United States Government’s first action regarding artificial intelligence regulations, the Safe, Secure, and Trustworth Artificial Intelligence” Executive Order. In what Biden calls a “landmark move….to ensure that America leads the way in seizing the promise and managing the risks of AI.”
Frankly, particularly in the world of copyright law, the legal standards keep shifting. The continuous rollout of new AI tools, as well as new applications of AI have created a scenario of continuously shifting sands when it comes to the intersection of law and the use of AI, that have resulted not just in confusion, but in an explosion of deep fakes. The use of AI-generated imagery and videos in the Israel-Hamas war have provided the most potent example, demonstrating how rampant and widespread the use of deepfakes have become.
According to the White House’s release, the EO “establishes new standards for AI safety and security, protects Americans’ privacy, advances equity and civil rights, stands up for consumers and workers, promotes innovation and competition, advances American leadership around the world, and more.”
While the Executive Order does not require specific action by Congress or state legislature in order to take effect, it does outline new standards for AI safety, and user security for AI-generated content. It is expected to be implemented over the next three to twelve months. While White House Deputy Chief of Staff Bruce Reed said in a statement that the EO represents “the strongest set of actions any government in the world has ever taken on AI safety, security, and trust,” the EO follows the European Union’s AI Act in an attempt to legislate the use of artificial intelligence.
The Executive Order (EO) reads as an effective example of strategic ambiguity, as it remains intentionally broad in scope. In general, the EO outlines 8 areas: consumer rights; jobs; equity & civil rights; innovation & competition; safety & security standards for AI; privacy protections; international leadership; and AI governance – with the aim of “protect[ing] Americans from the potential risks of AI systems.” Mostly notably: the EO proposes standards for watermarking AI-generated content, which will be regulated by The Department of Commerce.
In a word, no. More accurately, not quite.
An Executive Order carries the force of law to manage the operations of the federal government, but do not extend into the private sector. Based upon Constitutional authority (or congressional statutes), they directly impact US regulations, policies, and enforcement of laws. However, this is not legislation passed by Congress, and could be overturned or modified by future presidents. Executive Orders can also be challenged in court if deemed unconstitutional.
As Biden himself has stated, “This executive order represents bold action, but we still need Congress to act.”
Notably, this EO extends to US government agencies that produce digital content, stating that such content will need a watermark. It additionally requires corresponding text to include the voracity and authenticity of the original content.
“Generative AI” that creates “synthetic content”. The definitions of each are likely to only continue to grow. For the time being, it defines synthetic content as “images, videos, audio clips, and text, that has been significantly modified or generated by algorithms, including by AI”.
The Executive Order calls on Congress to adopt privacy legislation, but it does not provide a federally-mandated legislative framework.
Due to this lack of framework, it remains unclear how courts will interpret the EO’s directives, and what balance will be struck with existing consumer privacy & data rights statutes.
The EO requires that prior to the release of AI systems, companies must share safety test results with the U.S. Government. It also addresses training data for large AI systems, sets forth the need to evaluate how agencies collect and use commercially available data (particularly when that data involves personal identifiers), and prioritizes the National Institute of Standards and Technology’s development of standards for AI “red-teaming”.
It also emphasizes the necessity of standard and begins to define best practices for identifying, labeling, and authenticating synthetic content, in an attempt to create effective standards for Americans to be able to discern AI-generated content. Notably, the EO mandates that government agencies producing digital content must watermark AI-produced content, indicating the work’s authenticity.
The most interesting takeaway from this Executive Order: while it does reference intellectual property, it hardly provides any hint of a framework.
In a press release, Biden stated that this EO will broadly impact “investments in AI-related education, training, development, research, and capacity, while simultaneously tackling novel intellectual property (IP) questions and other problems to protect inventors and creators.”
However, the EO kicks back the intellectual property questions currently being tackled by both the Copyright Office and the United States Patent and Trademark Office. Section 5.2(c)(iii) directs the U.S. Patent and Trademark Office Director to provide recommendations to the President in consultation with the Director of the Copyright Office to determine potential executive actions that can be taken relating to copyright and AI “within 270 days of the date of the EO or 180 days after the U.S. Copyright Office of the Library of Congress publishes its forthcoming AI study.”
In a time when AI is exploding in popularity (and effectiveness), the future rulings based upon the parameters set forth in the EO could have a direct impact on online business owners in particular. For example, how much of a piece of content partially produced by AI constitutes “generated by AI”? (how do we define ownership, or what % counts?). Secondly, what in particular requires a watermark?
In particular, the lack of clarity on the myriad of questions around copyrights seems to indicate that for the time being, the federal government is placing the onus of determining intellectual property ownership questions on the courts (including the Copyright Office, and the USPTO). The Executive Order provides no clarification on the question of what percentage of a work (trademark or copyright) can be created by AI, and still fall within “fair use” laws. In short, this is just the beginning of what will be a rapidly-changing legal landscape- we are tracking the newest laws for you here.
LEAVE A COMMENT
View Comments