🔒At inaugural Hartford HealthCare AI summit, the key to the future is trust

As the use of artificial intelligence in health care continues to grow rapidly, one key factor putting the brakes on that growth is trust — especially when it comes to clinicians trusting the technology. That was just one of the messages delivered during Hartford HealthCare’s first AI summit, held both in person and virtually Tuesday evening. Called “The Transformative Power of AI: Accelerating Artificial Intelligence from Bench to Bedside Safely & Responsively,” the four-hour event attracted more than 300 people, including more than 200 watching online. It featured five presentations, including four panel discussions on AI topics including venture capital investments; the speed and safety of personalized care; the financial, legal and security risks; and the impact on patients. In his welcoming remarks, HHC President and CEO Jeffrey Flaks said that despite all of the recent negativity about health care — including federal cuts to Medicaid and research funding and complaints about rising costs — he nonetheless believes that this is “the best moment we've ever had in health care.” He believes that, he said, “largely because of what we're talking about today: There's never been an opportunity to get better faster. There's never been an opportunity to more creatively, impactfully create solutions for some of the greatest and longest-standing challenges that we've had in health care.” He added, “This is the best moment, and it's only going to get better.” How quickly that happens, though, will depend on educating patients and clinicians to create the trust needed for AI technology to grow. That’s according to Elad Walach, co-founder and CEO of Aidoc, which is developing AI-operated systems to help clinicians diagnose and treat patients. During a panel discussion entitled “Point of Care AI: The Reality of Safer, Faster, Personalized Healthcare,” Walach said that “the main limiting factor is trust.” “Look, the technology is not perfect, but it's still valuable and extremely useful,” he said, adding that “creating the trust, creating the explainability, is going to be really key as this evolves.” Panelist Dr. Padmanabhan Premkumar, a senior vice president with HHC and president of Hartford Healthcare Medical Group, agreed that trust is an important factor in AI development. He cited the work of K Health, which helped HHC launch a virtual health platform called Called HHC 24/7 in April. “That's a starting point for us,” Premkumar said. “That makes it easy for folks to see patients in a virtual care space. And then I think it makes it easy for us to come up with a differential diagnosis, for me to trust the AI is okay, but ultimately I'm still responsible.” He said he believes AI will ultimately improve access to care, as well as improve quality and safety. “So, I think that's where the next frontier for us is,” using AI to improve the data on patients, he said. “That augmenter of data that allows for our clinicians to function at the top of their license is the key to our success there.” Dr. Barry Stein, who moderated the panel, asked the participants whether the goal of AI is to automate functions or augment them. Walach said that when it comes to AI affecting clinical care, “we have to be extremely cautious.” He cited as an example an incident earlier this summer in which AI company Replit was working with a company to build a database, and the “AI decided to delete their code base and then lie about it.” After that incident, he said, “We have to realize that this is a new technology where we can’t even begin to imagine all the risks, and therefore, I am a very big believer, at least in the near future, of augmentation.” The panel’s third member, Ran Shaul, co-founder and chief product officer for K Health, said he believes AI is “very close” to providing automated clinical decisions. “Those machines are able to consume data and make decisions and learn from their mistakes very fast,” Shaul said, adding, “Humans are not going away, but they are going to adopt, to incorporate.” Other revelations from the event included a comment from Kevin Hamel, vice president of IT operations and technology platforms for HHC, who said the hospital gets “40 million email messages every 90 days, and 70% of those messages you never see because we strip them away.” He added that the majority of the emails “are malicious in some way, shape or form. We get probed all the time by foreign countries, but just had four incidents last week that my team was investigating in the span of one week. So, it's a very dynamic environment, and in some ways, it's a very dangerous environment, but there are plenty of safeguards that we can implement and reduce our risk to an acceptable level if we take the time to ask the right questions and understand how the technology works.”