As the role of AI in healthcare settings continues to evolve and generate debate, stakeholders from all around the industry are wondering what 2024 will bring.
Here are seven trends to watch this year:
1. EHR integration & collaboration
The idea of “co-pilot” tools are already gaining steam, with tech giant Microsoft joining forces with Epic to expand MyChart’s capabilities. Speaking at an event last August, Epic CEO Judy Faulkner foretold an AI tool that “will listen to the discussion” between clinicians and patients and streamline a variety of existing processes.
Industry stakeholders are watching these technologies with great interest. What use cases of Generative AI will Epic prioritize? What industry partnerships will form around the new tools?
2. Imaging as a litmus test
One healthcare space to watch is radiology, which might serve as a bellwether for industry adoption of larger multi-modal models ― those that blend text and images. Historically, radiology has been a litmus test for advances in medical technology. Machine learning and AI are no exceptions.
“Computer vision” is also buzzy. They have cameras in the OR so instead of scrub techs having to enter every step of a procedure in the EMR, the computer vision knows what and when and documents it automatically.
3. Flexibility with data
Its applications for radiology notwithstanding, the use of synthetic data has stirred some controversy. In general, synthetic data can mask bias and lacks the randomization inherent to real-life samples. For that reason, many reject AI-generated data as a poor substitute for the real thing.
At the same time, established healthcare companies need to remain flexible to stay relevant while remaining rooted in their core strengths. The industry is notoriously slow to change, and the use of generative AI is no exception. Whether linked to actual clinical encounters or not, generative AI has the potential to benefit patients and clinicians in ways we can’t imagine today.
4. AI and the prior authorization rule
The Centers for Medicare & Medicaid Services announced a new rule (CMS 4201-F) going into effect in 2026 that will require insurers to hasten their time to decisions.
There is ambiguity in the policy. Whether or not it inspires future policy changes will be revealed in time. Regardless of your position, the Prior Authorization rule highlights the ongoing need for shared, objective determination of medical necessity and data transparency ― both areas in which AI can play a valuable role.
5. Avoiding volatility
The pace of releases and advances in the generative AI space is hard to keep up with. Even organizations who are adept at keeping pace are learning it takes time to effectively implement any new AI initiative. Regulations on the use of AI are evolving concurrently ― more quickly in Europe than the U.S. ― and any early adopters must be mindful of this as well.
Those looking to invest in AI and adapt it to their organizations would be wise to go with tested, proven applications to avoid volatility. That will be essential to maximizing operational and financial ROI.
6. Think three steps ahead
Healthtech firms using AI need to be thinking three steps ahead at all times. The ripple effects of AI tools are inviting unforeseen consequences to the detriment of some very large stakeholders.
Take ChatGPT. The New York Times recently filed a lawsuit against tech giant Microsoft, complaining ChatGPT (in which Microsoft is the largest investor) is reappropriating the Times’ intellectual property to compose articles that compete against it. If a healthtech firms builds its own product on the backbone of an OpenAI product like ChatGPT, will clinical data be usable within its framework?
Concerns over data security and governance, technical maturity (particularly around generative AI tools), and preventing biases are better addressed sooner than later.
7. More resilient cybersecurity
Cloud-based security can improve the resilience of a healthcare organization’s cybersecurity response. If a ransomware attack happens, an organization can respond much faster in the cloud than with on-premises infrastructure, and will have a better chance of accessing protected backups.
The renewed need for cybersecurity vigilance isn’t unique to healthtech firms operating in the AI space, but their concerns are unique ― and growing. One recent report, the Google Cloud Cybersecurity Forecast 2024, warned that generative AI and large language models (LLMs) will be utilized in various cyber attacks such as phishing, SMS, and other social engineering operations with the purpose of making content and material (such as voice and video) appear more legitimate.
Photo: rudall30, Getty Images