At a Glance.
- US lawmakers introduce new bill centered on restricting AI model exports.
- Senate releases new AI policy framework.
US lawmakers unveil new bill to make it easier to restrict exports of AI models.
The News.
Late last week, a group of bipartisan lawmakers in the House of Representatives introduced a new bill, known as the ENFORCE Act, that would strengthen and simplify the Biden administration’s ability to impose export controls on artificial intelligence (AI) models. If this bill were signed into law, it would enable the Commerce Department to be able to bar Americans from working with foreign elements when developing AI systems that pose risks to United States (US) national security. This new legislation would specifically remove roadblocks that currently prevent the Commerce Department from regulating the exportation of open-source AI and give the Department the explicit authority to regulate AI systems.
The bill's sponsors include Republicans Michael McCaul and John Molenaar and Democrats Raja Krishnamoorthi and Susan Wild. This new legislation was crafted in collaboration with officials from the Biden administration.
The Knowledge.
Under current legislation, the Commerce Department has had difficulty regulating open-source AI models, which can be freely downloaded. However, with this new piece of legislation, both Congress and the Biden administration are looking to empower the Department to better enforce AI exports and make it more difficult for companies to legally challenge new export regulations.
This new bill comes after the Biden administration has increased its scrutiny over AI development and has been creating new export regulations regarding AI models. For example, last week, the Biden administration also updated rules to ban domestic companies from exporting proprietary or closed-source AI models to specific nations and actors. These updated regulations came after several other efforts were taken by the administration throughout previous months to enforce export controls related to both AI and other semiconductor chips. If this new piece of legislation were passed, the Biden administration would likely reveal new export restrictions given their involvement in creating this bill.
The Impact.
While this bill has yet to pass in either chamber of Congress, it is representative of rapidly growing concerns related to foreign use of AI. While it is unlikely that the average US citizen will notice the impacts of this new bill, the law could have significant implications for domestic AI developers. If this bill were signed into law, it would make it harder for AI developers to challenge new restrictions as well as could impact business dealings. AI developers should be aware of this new bill, its significant bipartisan support, and the implications it could have on existing projects and revenue streams. Additionally, developers should track this bill and any potential new regulations to ensure that they are in compliance with US law to avoid any unwanted legal complications.
Senate releases new AI policy roadmap.
The News.
On Wednesday, a group of bipartisan Senators released a new AI policy framework. The bipartisan group of Senators, led by Senate Majority Leader Chuck Schumer, aims to use this new framework to guide initiatives and legislative efforts over the coming years that will boost AI innovation and create new regulations. Senator Schumer is supported by Republican Senators Mike Rounds and Todd Young as well as Democrat Senator Martin Heinrich.
With this proposal, these Senators are calling for $32 billion in funding to be added by 2026 to fuel efforts related to managing AI. The roadmap’s priorities include ensuring the proper legal enforcement of existing AI laws while also advocating for the consideration of new laws. Additionally, Senator Schumer stated that the bipartisan group “supports a strong, comprehensive federal data privacy law to protect personal information.”
The Knowledge.
While this new framework has just been introduced and funding has not yet been allocated, the move is representative of the growing pressures facing the federal government to address AI. Not only does this framework aim to bolster funding but also Chuck Schumer mentioned how this funding aims to “cement America’s dominance in AI” and outcompete China. This latest federal move comes after months of both Congress and the Biden administration increasing their involvement in the development and use of AI. Over these months, various federal entities have released statements that they are either increasing funding for AI research, creating new guidelines, or restricting exports of AI-related technologies.
Notably, Senator Schumer’s comments regarding supporting a new federal data privacy law voiced critical support for the recently proposed American Privacy Rights Act, which was introduced in April 2024. This act was introduced and is supported in both the House and Senate and aims to create a new comprehensive federal standard for privacy across the US. With this new bill, this new federal privacy standard would focus on updating and improving the privacy of rights for US citizens in the modern world.
The Impact.
While this framework has only just been introduced and is not a piece of legislation, the impacts that this framework will have on the future of AI legislation could be significant. With the Senate Rules Committee already considering three bills related to regulating AI, this new framework will certainly have an impact on how these committees consider and propose amendments to each bill.
As the Senate continues to increase its involvement and supervision over AI development and usage, AI developers should understand that this document will likely greatly influence how the federal government approaches managing AI over the coming years. AI developers should understand this document and its implications to ensure that they are aware of the government’s potential impact on their process and to plan accordingly. These latest efforts being made by both Congress and the Biden administration demonstrate a consistent effort to address the existing gaps in both US privacy rights and AI development. US citizens should expect efforts similar to these to continue to gain more traction over the coming years due to the bipartisan support.
Other Noteworthy Stories.
US and China meet to discuss risks associated with AI.
What: American and Chinese officials met in Geneva to discuss risks associated with advanced AI technologies.
Why: On Tuesday, both American and Chinese officials met to discuss their concerns related to AI. These talks were focused on “an exchange of views on the technical risks of AI” according to a US official. During these discussions, the Biden administration raised concerns related to China’s use of AI in the military and other national security sectors. A US official commented that these talks were “not focused on promoting any form of technical collaboration or cooperation on frontier research in any matter” and that the US was adamant that “our technology protection policies [were] not up for negotiation.”
In a follow-up article, it was reported that these conversations were “candid and constructive” as delegates from both nations “exchanged perspectives on their respective approaches to AI safety and risk management.”
Microsoft deploys Generative AI system for US military.
What: Microsoft has developed a new offline generative AI model designed to handle top-secret information for the US military.
Why: Microsoft’s new military-centered AI system is based on ChatGPT’s GPT-4 AI system; however, it remains isolated from the internet and is only accessible through an exclusive US government network. This AI will be centered around handling top-secret information for US intelligence agencies.
Microsoft claims that this new AI system will be “clean,” meaning it can read files without learning from them. However, concerns have been raised about the potential for unintended leakage of sensitive information. This is the first time a major AI model has operated entirely offline.
McConnell opposes new legislation related to banning using AI to create deceptive election content.
What: Senator Mitch McConnell states that he will oppose a new bill related to banning the use of AI to create deceptive content regarding federal candidates to influence elections.
Why: On Wednesday, Senator McConnell voiced his disapproval of a new piece of legislation that would restrict the usage of AI. Senator McConnell stated that this new legislation “would tamper [with a] well-developed legal regime” that is used to take down false ads and would “create new definitions that could reach well beyond deep fakes.” Senator McConnell expressed concerns that this bill “could wind up barring all manner of photos and videos as long as the ill-defined ‘responsible person’ could deduce an alternative meaning from the content.”
The Protect Elections from Deceptive AI Act is supported by the Senate Rules Committee Chair Amy Klobuchar, Senator Josh Hawley, Senator Chris Cook, Senator Susan Collins, Senator Michael Bennet, and Senator Pete Ricketts. This new act would ban the use of AI to create misleading election content.