Biden’s AI order: 3 key takeaways
In an unprecedented move, the White House published the Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligenc e, on October 30, which opens with the warning: “AI has an extraordinary potential for both promise and peril”.
Dramatic, maybe. But after Joe Biden’s meeting with tech CEOs from Microsoft, Google, and OpenAI in May, plus hearing worrying testimony from experts, this is perhaps unsurprising.
The order comes as the US Patent and Trademark Office, the US Copyright Office, and the federal courts continue to grapple with complex IP questions raised by AI.
In Thaler v Vidal (2022), the USPTO and the Federal Circuit Court of Appeals ruled that an AI cannot be named as an inventor on a US patent, but left open questions regarding the patentability of inventions devised by people who used AI tools.
And earlier this year, the US Copyright Office retracted the copyright for the graphic novel Zarya of the Dawn by Kris Kashtanova, which comprises human-authored text combined with images generated by the AI tool Midjourney—drawing attention to the problematic crossover between human and AI authorship.
Here are three IP takeaways from the order worth noting…
- The USPTO must find answers to Thaler eligibility issue
The order emphasised the need to untangle the thorny issue of patent eligibility when it comes to AI, calling for the director US Patent and Trademark Office (USPTO), Kathi Vidal, to issue more concrete guidance.
While USPTO has been mulling this issue and has already issued a request for comments (RFC) seeking public comments on AI inventorship issues, this has undoubtedly ramped up the pressure on the agency.
As Robert McFarlane, partner at Hanson Bridgett, points out: “Alongside sweeping provisions meant to foster the safe and responsible development of AI, the order also includes provisions directed at furthering the development of intellectual property laws governing AI.
“This means that development of the law on the questions governing intellectual property protections for AI-generated subject matter will almost certainly accelerate.”
Among the specific IP-related provisions is a demand that the USPTO publishes guidance to USPTO patent examiners and applicants addressing inventorship and the use of AI in the inventive process within 120 days of the date of this order.
Further, the order stipulates that these should illustrate examples in which AI systems play different roles in inventive processes and how, in each example, inventorship issues ought to be analysed.
That’s not all. Within 270 days of the date of the order, the office must issue additional guidance to USPTO patent examiners and applicants to address other considerations at the intersection of AI and IP, including updated guidance on patent eligibility to address innovation in AI and critical and emerging technologies.
These provisions “may move the ball significantly forward on answering the questions left unanswered by Thaler”, predicts McFarlane.
As Blair Jacobs, principal at McKool Smith, observes: “The IP industry will likely continue to analyse and consider the patent eligibility of AI and the order directs consideration of this issue as well.
“It may be that eligibility law as a whole needs to be straightened out before the complexities of AI-based inventions eligibility can be resolved.”
- ‘Hardline’ courts can expect patentability guidance
Additionally, the order may be the starting point for a much-needed framework for the courts.
Jacobs explains that, to date, the Federal Circuit has taken “a hardline view” on the issue of patentability of AI “based largely on dated language from the Patent Act concerning ‘persons’ creating patentable subject matter”.
Consequently, a big question mark looms over certain scenarios that could lead to AI contributing to patentable subject matter.
“Guidance on inventorship issues related to using AI as a tool,” adds McFarlane, would provide a framework for the courts to analyse contributions of AI tools to the inventorship process, and ideally, “create a minimum human contribution that will qualify for human inventorship”.
Such a development, he says, would go some way towards improving the consistency in how the USPTO addresses inventions created with the use of AI.
And even though the federal courts are not bound by USPTO guidance, McFarlane notes that— as patent applications are issued or denied based on these new guidelines—federal district and appeal will be provided “with clear and detailed records from which to develop further law on these critical issues”.
- The era of AI policymaking has begun
If the order’s mandates are realised, it will certainly keep lawyers busy.
Notably, the order asks that the USPTO director consult with the US Copyright Office director and provide recommendations for further executive action, including potential copyright protection for works created using AI and the treatment of copyrighted works in connection with training AI models.
As Jacobs explains, such novel developments—if fully realised—would demand that IP practitioners “think strategically and creatively for years to come”.
Dion Bregman, partner at Morgan Lewis, agrees that the order means that lawyers and in-house legal departments will need to keep an even closer, razor-sharp eye on developments related to AI.
“Lawyers, especially those working in related areas such as IP, will need to monitor these developments and understand how they impact their practice areas.”
But, he cautions, Biden’s order has yet to be scrutinised by an increasingly politically fractured, and frequently deadlocked, Congress.
“The order’s breadth and depth represent a powerful statement by the president but is constrained by the rulemaking system and requirements that undertakings be authorised by congressional appropriation.”
“It is, therefore, a policy roadmap,” he concludes, “but nothing has yet changed from a regulatory requirements perspective”.
