At the end of last month, President Biden issued what was described as a “landmark” Executive Order on “safe, secure, and trustworthy artificial intelligence.” The Order is intended to “ensure that America leads the way in seizing the promise and managing the risks of artificial intelligence” by establishing new standards for AI safety which “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.”
“As part of the Biden-Harris Administration’s comprehensive strategy for responsible innovation, the Executive Order builds on previous actions the President has taken, including work that led to voluntary commitments from 15 leading companies to drive safe, secure, and trustworthy development of AI,” the Order reads in part.
As I mentioned earlier today, there are questions aplenty regarding the ethics and safety of artificial intelligence. President Biden’s Executive Order is proof positive of that. Nonetheless, it’s also important to simultaneously recognize how AI can genuinely benefit humankind such as its applicability towards helping the disability community live better.
One example is the work Punit Soni and his team at Suki AI is doing to make electronic health records more accessible to doctors everywhere. The Bay Area-based healthcare tech startup builds software that uses artificial intelligence to make documenting in patients’ charts easier and more efficient. Soni, a former product manager at Google, spoke with me in mid-May about his company’s mission to make healthcare tech “invisible and assistive so that clinicians can focus on patient care.”
The idea that EHRs, shorthand for electronic health records, can be made more accessible transcends sheer convenience. As I wrote in the spring, utilizing AI in this way can be a boon to doctors who have health conditions all their own. It’s entirely plausible someone may not be able to spend lots of time staring at a screen or holding a pen and paper in order to make notes in someone’s chart. That the process can be computerized becomes not only matters of convenience and efficiency but, as ever, it also becomes a matter of—what else?—accessibility.
Given AI’s relevance to accessibility, as well as my prior coverage of Suki AI, Soni and I reconnected last week over email for a brief interview about the aforementioned Executive Order. He told me the Order is “a much-needed step [which] demonstrates that AI is front and center for the administration.” Although the Order sets forth principles that Soni “should be material for AI companies” like his, it isn’t without its flaws.
“The technology of today will be the table stakes of tomorrow,” Soni said.
“Regulations requiring notification steps based on the size of AI models is not useful, and carries the risk of creating onerous requirements, especially for smaller companies,” he said. “What should be considered is an AI version of HIPAA, where there is detailed guidance on data usage, infrastructure, training data, and how to tackle bias, model leakage, and other factors in a way that promotes safety, national security, and equality while preventing regulatory capture by large companies. It’s imperative to have industry representation in these committees and task forces that are representative of the vibrant startup AI ecosystem, not just those who can afford lobbying power.”
Soni called Biden’s Executive Order “timely and material to Suki and to our ambitions” insofar as the company believes AI will “manifest itself as an assistive layer that sits on top of all of healthcare [and enable] more efficient and democratized access for everyone.” Moreover, he said the AI-based assistant for providers will ultimately have “a meaningful impact” on Suki’s business and the tech industry writ large. Startups and small companies do well to drive technological innovation and boost the economy, but are less resource rich than much larger, incumbent organizations. As such, regulations must “consider how to balance requirements without creating overhead that puts small companies at a disadvantage.” To do so would “create a significant barrier to innovation and could ultimately lead to only a few large companies that can commercialize the technology because they have the resources to do so.”
“Suki strongly recommends a task force [and] organizing group that has representation that includes impactful startups along with industry incumbents and other constituencies so that influence on the impact of this [Executive Order] is not constrained to those who have lobbying power,” Soni said of how to best regulate AI with lawmakers.
When asked about how prudent it is for politicians to legislate technology—who may not necessarily understand all the intricacies about how a piece of tech works—Soni was diplomatic in his response. He said it’s important to “include all relevant stakeholders in designing policy” which include people like founders, academics, end users like (in Suki’s case) clinicians, and more. Doing so, he added, allows for “a broad perspective that accounts for the full scope of benefits, drawbacks, and considerations” which brings a level of pragmatism to regulation.
“Companies like Suki are a necessary, but not sole part of, that [regulatory] ecosystem,” Soni said.
Read the full article here