Is A.I. Scary or Exciting? – YES!
Let me start by saying I love Chat GPT, but A.I. is more than just chat-bots. I was floored during last year’s Canadian Organization for Rare Disorders (CORD) conference where I learned about other A.I. tools in diagnostic services and in drug production. I have read some fantastic stories about A.I. triage tools for the Emergency Department, and a deeply troubling story about an A.I. case management tool failures in Mental Health clinics. I admit to having a very limited understanding of A.I. but I do believe this technology is here to stay. Health and social service sectors cannot succumb to Ostrich Syndrome, burying our heads in the sand, while we let other sectors drive the conversation about these tools. In an environment with increasing pressures to deliver quality quickly and with fewer resources, we will need to maximize strategic A.I. adoption while simultaneously being aware and evaluating the risks. How do we as a system respond to Canada's voluntary code of conduct relating to advanced generative AI systems?
A.I. will help reduce administrative burdens, reduce time and errors in diagnosis (eventually/hopefully), provide more efficient case management, and supercharge our research practices. AI is exciting.
AI can make mistakes, create mis/disinformation farms, and open deep ethical and privacy concerns. AI is Scary.
As a sector we need to be agile and resilient enough to hold these contradictions simultaneously. I will continue to use Chat GPT to help with my creative processes and to meet production expectations. I will also be critical of what it generates, ensure I am not relying on it as a source of truth, and always further engage with it’s output.
I find too many leaders are afraid to admit they’ve used this tool publicly. Perhaps worried it suggest their ideas aren’t their own, or that they aren’t working hard enough. Generative A.I. is a tool that we need to normalize in our work settings, failure to do so will needless hold us back and discourage the next generation of our workforce (who will be infinitely more capable with these tools than we are) to perform at their best! I almost always have my A.I. tab open while I’m writing these days.
“The Medical AI Podcast: The Application of Chat GPT to Healthcare” discusses A.I. use in pediatrics in the NHS. The story they shared was of a child being discharged and their parents were handed the traditional discharge package. Buried in the tome of information was the importance of medication compliance. NHS staff were able to run this discharge package through an A.I. tool asking it to synthesize the medication compliance bits into a narrative story for a grade 4 child. They then ran that story through an image generator (the likes of Dall-E) and instantly had a colouring book for a 9-year-old about medication compliance – a process that would take maybe half an hour. Story’s like this make me confident A.I. can be a powerful tool in maximizing impact with limited resources.
So for those who haven’t tried Chat GPT yet, I asked it to write the next section of this article based on the prompt: “Types of AI in Healthcare and Social Services with pros/ cons.” (This section is copy pasted so please put on your critical evaluation hat and let me know if it says anything you find worrisome!)
____
Diagnostic AI: One of the most promising applications of AI in healthcare, diagnostic AI leverages machine learning algorithms to analyze medical data, such as X-rays, MRIs, and patient records. It can help identify diseases and conditions faster and with high accuracy.
Pros: Rapid diagnosis, reduced human error, and the potential for earlier intervention.
Cons: Limited access to advanced diagnostic tools in underserved areas, concerns about data privacy, and the need for human oversight to validate results.
AI in Drug Discovery: AI accelerates drug discovery by analyzing vast datasets to predict potential drug candidates and their efficacy.
Pros: Speeding up the drug development process, identifying novel therapies, and reducing research costs.
Cons: Ethical concerns in patenting AI-generated drug formulas, the need for rigorous clinical trials, and the risk of overlooking traditional methods.
Administrative AI: AI-driven tools streamline administrative tasks in healthcare and social services, such as appointment scheduling, billing, and record-keeping.
Pros: Increased efficiency, reduced administrative costs, and more time for healthcare providers to focus on patient care.
Cons: Potential job displacement, data security risks, and the challenge of ensuring AI systems align with regulatory requirements.
AI for Predictive Analytics: AI models analyze patient data to predict disease outbreaks, hospital resource allocation, and patient readmissions.
Pros: Improved resource planning, better patient outcomes through timely interventions, and cost savings for healthcare systems.
Cons: Data biases, privacy concerns, and the need for continuous refinement of predictive models.
AI in Social Services: AI assists in social services by automating case management, identifying at-risk individuals, and optimizing resource allocation.
Pros: Enhanced case management, more targeted interventions, and improved allocation of social services resources.
Cons: Ethical dilemmas related to automated decision-making, potential reinforcement of existing biases, and the importance of human empathy in social services.
____
A.I.'s integration into our healthcare and social services systems marks a pivotal moment. It is as exciting as it is scary. I invite leaders to talk openly and often about these tools so that healthcare and social service sectors can be part of the conversation in how we can leverage this technology in safe, effective, and innovative ways while reducing risk. As governments discuss how to navigate A.I. policy it will be important for health leaders to be at those tables. I believe we must be courageous in its adoption, and cautious in it’s implementation.
What do you think? Agree, Disagree? Let’s start a conversation!
