Category: Artificial Intelligence

Study: AI is in the Eye of the Beholder

Posted on by Jason Barry

At a time when new healthcare AI solutions are getting unveiled every week, a study in Nature Machine Intelligence found that the way people are introduced to these models can have a major effect on their perceived effectiveness.

Researchers from MIT and ASU had 310 participants interact with a conversational AI mental health companion for 30 minutes before reviewing their experience and determining whether they would recommend it to a friend.

Participants were divided into three groups, which were each given a different priming statement about the AI’s motives:

  • No motives: A neutral view of the AI as a tool
  • Caring motives: A positive view where the AI cares about the user’s well-being
  • Manipulative motives: A negative view where the AI has malicious intentions

The results revealed that priming statements certainly influence user perceptions, and the majority of participants in all three groups reported experiences in line with expectations.

  • 88% of the “caring” group and 79% of the “no motive” group believed the AI was empathetic or neutral – despite the fact that they were engaging with identical agents.
  • Only 44% of the “manipulative” group agreed with the primer. As the authors put it, “If you tell someone to be suspicious of something, then they might just be more suspicious in general.”
  • As might be expected, participants who believed the model was caring also gave it higher effectiveness scores and were more likely to recommend it to a friend. That’s obviously relevant for those developing similar mental health chatbots, but a key insight for presenting any AI agent to new users.

An interesting feedback loop was also found between the priming and the conversation’s tone. People who believed the AI was caring tended to interact with it in a more positive way, making the agent’s responses drift positively over time. The opposite was true for those who believed it was manipulative. 

The Takeaway

The placebo effect is a well documented cornerstone of medical literature, but this might be the first study to bridge the phenomenon from sugar pill to AI chatbot. Although AI is often thought of as primarily an engineering problem, this research does a great job highlighting how human factors and the power of belief play a huge role in the perceived effectiveness of the technology.

Bain & Co: Getting the Most Out of Generative AI

Posted on by Jason Barry

Bain & Company is back at it again with more generative AI research, this time offering a series of ways for providers to get the most out of the tech without falling into potholes of hype.

The in-depth report gives a comprehensive overview of the current generative AI landscape, and delivers solid insight into the priorities of health system executives (N=94): 

  • Top use case priorities (next 12 months): charge capture & reconciliation (39), structuring & analysis of patient data (37), workflow optimization (36). [Chart 1]
  • Top use case priorities (2-5 years): predictive analytics & risk stratification (44), clinical decision support (41), diagnostics & treatment recommendations (37). [Chart 2]
  • Biggest barriers to implementation: resource constraints (46), lack of technical expertise (46), regulatory & legal considerations (33). [Chart 3]

Start small to go big. Although the survey itself included some valuable stats, the spotlight was stolen by Bain’s particularly pragmatic framework for guiding new implementations. 

  • Pilot low-risk applications with a narrow focus. Bain found that the systems already seeing the most success with generative AI are testing solutions in low-risk use cases where they already have the right data and can create tight guardrails (chatbot support, scheduling, rev cycle).
  • Decide to acquire, partner, or build. Bain recommends that CEOs think about different use cases based on availability of third-party tech and importance of the initiative.
  • Funnel experience into bigger initiatives. As generative AI starts to mature, organizations that gain experience and strategy alignment today will be best positioned for the more transformative use cases once they become clear.
  • Generative AI isn’t a strategy unto itself. Bain found that the trait separating top CEOs is their discipline, ensuring that every generative AI initiative reinforces their overarching goals as opposed to implementing shiny bells and whistles.

The Takeaway

It’s easy to get caught up in the generative AI hype cycle, so it was refreshing to see Bain recommend the one-foot-in-front-of-the-other approach to new implementations. Nearly every hospital boardroom is debating a massive list of potential AI investments, and although the home run use cases will be here soon, the consensus strategy for getting on base seems to be making low-risk plays with an immediate impact.

Hippocratic Emerges to Bring LLMs to Healthcare

Posted on by Jason Barry

Although we touched on Hippocratic AI’s emergence from stealth last week, the startup struck all the right chords with its talk of generative AI and large language models so we’re doing a deeper dive to unpack the hype. 

Hippocratic AI debuted with $50M from a massive seed round co-led by the VC power duo of General Catalyst and a16z, giving the company a “triple-digit millions” valuation right out of the gate.

  • The company’s mission is to transform healthcare through the power of “safety-focused” generative AI, but potential use cases are still in the works, with early ideas revolving around consumer-facing tasks like diet planning and medication reminders.
  • The founders even told STAT that their goals for the funding are only exploratory: developing a LLM that’s fine-tuned for healthcare, heavily testing it against knowledge benchmarks, then measuring its bedside manner. 

If including Hippocratic in the name wasn’t enough of a hint, pressure-testing the model’s accuracy is in the company’s DNA, and it isn’t planning on rushing into clinical care.

  • Accuracy is ensured through reinforcement learning from human feedback (RLHF) performed by medical professionals – a strategy that’s apparently outperforming GPT-4 on 105 of 114 healthcare certifications. [Comparison Chart]
  • As for bedside manner, Hippocratic is planning on “detecting tone” and “communicating empathy” better than rival models, and developed its own benchmark for behaviors such as “taking a personal interest in a patient’s life.”

The Takeaway

Healthcare hasn’t always been kind to AI-first businesses, with IBM having to let go of its Watson Health division and Babylon recently meeting the end of its road as a publicly traded company. That said, both of those examples paddled too early to catch the current generative AI wave with its mile-long barrel of new tech and excitement. It’s too early to tell whether Hippocratic will buck the trend, but if there was ever a moment to try it – this is it.

Get the top digital health stories right in your inbox

You might also like..

Select All

You're signed up!

It's great to have you as a reader. Check your inbox for a welcome email.

-- The Digital Health Wire team

You're all set!