5 Predictions for Healthcare AI in 2026

Barry Stein, Hartford HealthCare’s Chief Clinical Innovation and Medical Informatics Officer and founder of the Center for AI Innovation in Healthcare, says he has a good idea where AI will be going this year.

AI is poised to do big things in healthcare this year, according to Hartford HealthCare’s Barry Stein, MD, MBA. The key will be in moving beyond the hype and actually making an impact on clinical care.

Stein, the health system’s VP, Chief Clinical Innovation Officer and Chief Medical Informatics Officer, as well as founder of the Hartford HealthCare’s Center for AI Innovation in Healthcare, says healthcare leaders need to see AI as part of an ecosystem. And it will only work in healthcare if it all the right safeguards are in place.

“We approach AI in the same way as we approach a new device, a new robot, or a new drug,” he says. “People think just because it's software, it doesn't have the same potential impact as a new drug or a new robot or a new device. It has as much upside. If you can leverage it to the fullest and you know how to use it, and it has equal amount of downside if you don't know how to use it and you give it to the wrong person and you don't know what's going on. So with a new device and a new drug, you have to know how it works.”

Beyond that plea for governance, Stein says AI has the potential to right the wrongs of past technology integrations and once again give clinicians and their patients the time and space to talk.

Stein offers 5 predictions for AI in healthcare in 2026. They are:

• AI will make care more human, not less. From conversational AI to foundational imaging detection models to algorithms for nurse scheduling, he says, AI will bring clinicians closer to their patients and remove the barriers that past technology implementations have raised.
• AI becomes a true clinical partner. “Driven by consumer expectations for immediacy and personalization, AI will increasingly improve care coordination, enhancing the experience for patients and clinicians,” Stein says.
• Ambient AI becomes vastly more context-aware. “Rather than only capturing conversations, AI will also summarize a patient’s chart for the provider before the visit; and during the encounter create a pertinent summary, flag any payer requirements, create orders for studies and procedures, obtain necessary prior authorizations, and generate a visit summary while the patient in the room,” he notes.
• 2026 is the year of trust and transparency. “AI is not just documenting care but doing so in a way that patients can understand and clinicians can trust,” Stein says. This particular topic is likely to gain a lot of attention, as states vie with the federal government to govern the technology and create guardrails. Others have said that 2026 might see the first lawsuit brough against a healthcare organizations for not using AI to improve care.
• Patients choose care sites based on AI experience. “Patients will come to value the improved experience that AI makes possible as a differentiator and choose where to receive their care based on how AI benefits them with improved access and convenience, and their confidence in the quality and safety of care,” Stein says. 

Stein says Hartford HealthCare has six pillars for AI development: (1) An ecosystem of partners, (2) research, (3) governance, (4) education, (5) change management and (6) data integration. Alongside that are four strategic imperatives that every new AI tool or program must address: (1) quality and safety; (2) access, (3) affordable healthcare, and (4) care equality.

That’s a lot of checks to mark off. But Stein says they’re essential for managing a technology that continues to evolve.

Barry Stein, MD, MBA, VP, Chief Clinical Innovation Officer and Chief Medical Informatics Officer at Hartford HealthCare and founder of the Center for AI Innovation in Healthcare. Photo courtesy Hartford HealthCare.

“How do you monitor that it's performing the same way as you expect it to on the day you deployed it?” he asks. “The bottom line is these things are highly adaptive, and they’re changing all the time. And we as a society … have not yet figured out how to monitor it.”

When Stein talks of an AI ecosystem, he envisions a platform built on data, the fuel of every single AI tool and program. And he’s not alone when he says the healthcare industry has to understand data management before it can truly move forward with AI. Healthcare has access to more data than ever before, but healthcare leaders have to know how to gather, assess, use and protect that data.

Because healthcare providers work with that data every day, he says, they have an advantage over companies outside the industry that are looking to compete with health systems and hospitals. There’s a clear difference between addressing a consumer’s needs and addressing a patient’s needs.

“If you want to penetrate healthcare, you've got to have research,” he says. “You've got to have data. You can't just come into healthcare and say, ‘You know what, this stuff is pretty cool. Why don't you try it?’ You know it doesn't work like that. You've got to have data, real research data.”

As well, Stein says the perception within and outside the industry that AI will replace clinicians is misguided. AI needs to be designed to support clinicians by reducing the barriers that technology has put between them and the patient.

“This is not AI operating autonomously,” he points out. “This is AI complementing clinicians, allowing clinicians to do what they're trained to do, remaining connected to their patients. Give me the most important information. I'm remaining critical in the loop. I'm still overseeing everything, but I've become hyper-efficient, and for the patient [I’m creating] a much more pleasant experience.”

The argument that AI will replace doctors and nurses, he says, is wrong. AI will take over the tasks that doctors and nurses shouldn’t be doing, and it will be the tool that patients use to get closer to their doctor, nurse or care team.

AI should “take work out of the workflow” for clinicians, he says.