Artificial intelligence (AI) has garnered significant attention, particularly with the rise of generative AI, which captivates us with its ability to create content, simulate human-like behaviors, and generate novel outputs – while at the same time instilling fear of a Terminator-like future. And while generative AI gets the press, it is adaptive AI, working in the background, that is truly driving sustainable value across industries such as healthcare.
Generative AI, like ChatGPT, is a type of predictive AI that uses neural networks to support large language models trained from big data sources like the internet. The models run in response to text queries that are used by the AI to generate content, images, text, and even music. This simulates human creativity, enabling autonomous creation without human intervention. However, it can lack controllability and consistency, and at times, raise ethical considerations.
Adaptive AI on the other hand runs predictive AI models triggered by real-time events. Adaptive AI can orchestrate workflow and learn from its interactions, continually evolving and improving performance. This narrow use of AI adjusts to new information, making it more flexible, precise, and applicable in diverse scenarios. It is especially useful for automation, real-time decisioning, and orchestrating proactive consumer experiences across complex processes.
While generative AI captures our imagination and emotions, it is adaptive AI that creates tangible business value with its ability to learn from experience, adapt to changing environments, and optimize performance to enhance efficiency, accuracy, and decision-making in industries like healthcare.
Here are just a few ways adaptive AI is impacting healthcare:
- Supporting call center agents using adaptive AI co-pilots to “listen in” to conversations using a combination of adaptive AI and natural language processing (NLP). The AI presents information in real time to help agents fill out forms, find, filter, and present information (e.g., benefit or claims status). It guides the agent through the conversation, supporting compliance, providing coaching tips, and creating a better experience for the caller and the agent. After the call is finished, the AI continues to work summarizing and creating a call wrap-up report and initiating follow-up actions. This use of adaptive AI drives tangible business value by onboarding agents quickly and reducing training costs and compliance risks – all while improving first-call response, call handle time, and user and member satisfaction.
- Automating complex healthcare processes that engage people (e.g., patients and caregivers), providers, and payers requires making decisions at critical points to achieve the best outcomes. This is an ideal use of adaptive AI. The flow of information is often siloed in many systems, which increases friction that drives inefficiency, waste, and cost. Additionally, patients, providers and payers engage infrequently through occasional visits and calls. This increases the “blind spots” in the data record where providers and payers don’t know what’s happening with the patient. Now adaptive AI is being used to proactively engage patients, continually analyzing, prioritizing, and routing work, along with adding actionable insights to the data record while following complex processes to ensure that tasks get completed and people don’t fall through the cracks.
- Activating patients and caregivers to provide the necessary skills, knowledge, and preparation to successfully manage their own health. Activated people have better health and outcomes as well as lower healthcare costs than people that are not actively engaged. But achieving this requires more than providing educational content during brief encounters – often presented during moments of stress like a hospital discharge. The reality is that learning requires repetition. Humans must repeat and reinforce concepts to turn information into knowledge and habits. Adaptive AI provides an innovative approach to creating this level of engagement to turn educational content into a learning experience. Imagine a chatbot sends a prompt to remind or educate (e.g., watching for signs of infection is critical after your discharge), or asks the person to observe and respond (e.g., Do you notice any signs of infection? Yes or no?). When prompted over time, this simple adaptive AI interaction provides repetition to help patients better retain knowledge. By taking an action like observing, the concept is reinforced. The response provides insight into the person’s situation at that moment. When repeated over time, adaptive AI continuously assesses the person’s status.
This ability to continuously engage a large population of individuals to assess status, anticipate needs, and act before they become bigger problems addresses one of the biggest issues in healthcare: The fact that patients are responsible for most actions – not the doctors, nurses, hospitals, payers, or care teams. This effectively gives the most critical actions to the most vulnerable point in the healthcare system.
Using adaptive AI to monitor, assess, guide, and orchestrate actions (e.g., post-surgical discharge) offers an innovative new way to address the problem. By using AI in the background, running models to constantly learn and adjust to changes in the person’s journey, adaptive AI can be used to educate, remind, and assist patients and consumers in taking critical actions. The AI then orchestrates activities that can achieve the best outcome. - Activating communities is another area where adaptive AI will play a significant and valuable role. Many of our current health problems are caused by underlying social factors, such as financial and health equity, an aging population, isolation and loneliness, smaller families, behavioral and mental health issues, and many other factors. These problems can’t be solved with a perfect health system – it just serves as a backstop for problems that originate in a social system that’s broken for too many people.
Using adaptive AI offers a way to effectively turn the local community into a virtual family unit, identifying needs, distributing work, and making sure that critical actions are taken. Using AI to engage the local community of volunteers, community-based organizations, social workers, community health workers, neighbors, religious organizations, and others through a myriad of digital channels offers a proactive way to direct independent activities to the most critical people – effectively triaging the population to optimize the use of limited resources.
In this way, adaptive AI will be a key element in creating future job opportunities that enable humans to address the social needs and gaps that impact our unsustainable cost of healthcare. For example, using adaptive AI to equip volunteers to check in on neighbors decreases social isolation and offers a way to monitor the needs of those patients and consumers who don’t connect digitally, are housebound, or are elderly and isolated.
The value of activating communities using adaptive AI will be key to realizing healthcare’s Quintuple Aim goals of improving patient experience, achieving population health success, and team well-being to drive cost reduction and improve health equity.
Reducing the risk of AI
While both generative and adaptive AI have transformative potential, ethical concerns arise – especially regarding generative AI's potential misuse for deepfakes, misinformation, and biased outputs. Adaptive AI, while less susceptible to such risks, also requires careful monitoring to avoid reinforcing biases and ensuring transparency in decision-making processes in its real-time decisioning. This demands a way to monitor adaptive AI models to ensure ethical use as well as audit to the level that explains the outcomes of the AI models in human terms.
Challenges, such as data privacy, interpretability of AI decisions, and scalability, persist. However, embracing adaptive AI's continuous learning capabilities presents significant opportunities for enhancing productivity, personalization, and decision-making.
As AI continues to evolve, the fusion of generative and adaptive AI holds promise for groundbreaking advancements. However, prioritizing the responsible and ethical development of AI remains paramount for its widespread adoption and societal benefits.
In conclusion, while generative AI captures headlines, it is adaptive AI that serves as the backbone, driving tangible value across sectors like healthcare. Its ability to adapt, learn, and optimize based on experience positions it as the driving force behind AI's transformative potential in various industries, paving the way for a future where innovation and practicality converge for the betterment of society.