ChatGPT and other large language models, such as those behind ChatGPT, are already being used in healthcare. Developers are diving in to the deep end of a pool to create potentially time-saving tools that can be added to their main technology. Providers are using the tech to do everything from patient engagement and clinical decision support.
Is generative AI ready to take on the world? What can it do at this stage? What are the key issues for health IT leaders, workers in provider organizations and other stakeholders?
Healthcare IT News We sat with Vrinda Khurjekar, senior director for Americas at Searce AI consultancy to get her expert responses to these and other questions.
Q. What do you think of generative AI (artificial intelligence)? Can it be used in healthcare today?
A. After the internet revolution, I think gen AI will be the next big technological advance that will fundamentally change the way we live and work. The generative models have been around for some time, but their mainstream adoption will be what drives these transformations on a large scale.
The applicability of the technology across industries is already evident, and healthcare is no exception. It’s not just about improving the patient experience. It can also be used to improve the efficiency of the hospital care system.
It is still early days, and many things are yet to be revealed – for example, which use cases cause the least hallucinations [errors] The results are promising. We will need to define data privacy and sharing regulations to ensure that the models are ethical.
It is unlikely that it will be able to be used reliably for every possible use case. But can we get to that point? I would say yes, with the right governance, and controls.
Q. What are the main challenges that generative AI faces in gaining mainstream adoption?
A. There are many:
-
Lack of clarity about where to apply. Businesses are choosing non-value adding use cases because of the urgency to adopt gen AI. We hear conversations like “We need something with gen-AI, our board wants that” and when asked what exactly, there are many instances where the answer is unclear. We will tell clients that in order to do X, we need your data to be in order. Customers say, “Well, this is not a high priority.” We can’t do that.
-
Hallucinations. Most of these models are hallucinations. It will take some time to stabilize them for the relevant business functions. Many businesses are not willing to wait for the bugs to be ironed out.
-
Reduced line of sight to ROI The cost of implementing these models is high because of the shortage of talent. Companies are struggling with how to justify a large up-front cost for a long-term ROI. Access will be limited until economies of scale can be reached. This will limit mainstream adoption.
Q. Many health IT companies, including electronic record companies, are embracing generative AI functionality. What should hospital executives and clinicians do when faced with new AI tools that generate data?
A. Establish gen AI checks into your vendor onboarding. Evaluate the security, compliance, and governance policies of your vendors. It is almost necessary to rewrite internal guidelines in order to accommodate the new nuances that AI technologies bring. Also:
-
Have zero trust policy. You will be subjected to stricter compliance audits regarding what data is shared between IT providers and how you are able to see it.
-
Invest in internal experts Everyone is learning and adapting to the new space. In order to be successful, you will need to invest in internal experts as well as partner with the right industry partners that can guide you through the evaluation and implementation processes of these vendors.
Q. Oracle made waves recently with the launch its Q. Clinical Digital Assistant generative artificial intelligence. Please tell us a little bit about this tool, and why you believe it is the best way forward at this point.
A. It was exciting to see Oracle introduce a set generative AI services within its EHR solution portfolio. Oracle’s vision for the Oracle Clinical Digital Assistant, which will streamline administrative tasks using voice commands, will be a game changer in the way clinical visits will be conducted in the future.
We will need to see how the features are actually implemented in the next 8-12 months. I am convinced that the investment has been made in the right areas. You want your doctor to be well prepared for your visit. They should spend less time reading reports and other documents.
Patients should also be able to access as much information as possible about their conditions, causes etc. This digital assistant will not only save time for administrators and clinicians, but also help maintain a seamless patient experience.
Q. What role do you envision generative AI playing in healthcare in five years?
A. Five years is a long time, especially when we see new applications almost daily. In five years, I believe the hype will have subsided. We will see a lot of business processes that are redesigned when gen AI becomes mainstream. There will be a lot of newer technology and solution providers that will dominate niche markets.
We will see a complete revamp of the compliance and regulation guidelines to monitor bad actors, as well as protect individual rights.
Most importantly, we will see a shift in the world towards a lot more preventive care, helping us to proactively monitor our health, which will, in turn, improve the life expectancy for every individual.
Follow Bill’s HIT reporting on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is published by HIMSS Media.
You can apply Generative AI to almost any healthcare use case that you can imagine posted first on https://www.twoler.com/
No comments:
Post a Comment