The Challenges of AI
Tanya Procell, RN, president of clinical services at Provider Professional Services, said, “We see more and more tasks integrated into our workflow that use AI.” She observed that nursing staff generally appreciate the value of the technology and understand that it is there to support them, not to replace them. However, she said that their comfort level must come with some cautionary measures. For instance, she observed, “You have to be very careful about what kind of AI you’re using and that you’re not exposing HIPAA-protected information, so we caution against using ChatGPT or similar AI that’s not integrated into your health record.” Holahan added, “It’s always good to do your homework. There are a lot of AI tools out there. Some are better than others.” It is important to know, for example, if the AI tool was developed for clinical efficiency or insurance coverage and how it accounts for biases.
It’s also important to remember that AI is only as accurate or effective as the data that goes into it. It is possible to feed in information that is designed to produce a particular conclusion, such as a treatment algorithm that eliminates certain groups of patients from coverage. It is key, said Holahan, to maintain person-centered care and individualize treatment and care plan decisions for each person. Team members, as well as residents and families, need to understand the appeals process for any coverage denial and how to ensure everyone gets the care they need.
Additional Challenges
Environmental impact
Large amounts of power, energy, and water are needed to operate generative AI language models.
Enhanced surveillance
AI-enabled cameras on personal devices, in buildings, in organizational systems, and elsewhere pose a potential threat to privacy.
Data manipulation
As stated earlier, depending on the data fed into ID, it is possible to influence or skew results.
Potential for intellectual property theft
Data and information, including writings and artwork, can be used or taken by AI without crediting or paying the originators. Concerns about this have led many scholarly journals to prohibit the use of AI for writing, editing, or reviewing articles.
Misinformation
The ability to enable and promote the spread of misinformation, making it more challenging to distinguish accurate information and images from falsities and deep fakes.