The buzz surrounding artificial intelligence is deafening. AI has crash-landed in the zeitgeist. It’s invaded art, design, science, technology, business, and healthcare. The promises of AI are plentiful: productivity efficiency, decision-making, automation, and generally making life better for everyone. The talk of human workers losing their jobs as machines replace them is equally plentiful.

There is a lot to process. Meanwhile, healthcare revenue stakeholders must make critical decisions about how and where to invest in their business to improve productivity and increase margins; do they invest in people, technology, or both?

A new Sage survey asked about 2024-2025 healthcare C-suite priorities1. 57% of survey respondents said growing revenue is their top strategic initiative, 55% say staff recruitment and retention, and 46% say reducing costs. So, people and technology will both be essential investments going forward.

If technology is part of your roadmap, you will inevitably look at solutions that feature AI. Does AI deliver on its promises and make a tangible impact on the business? The answer depends on what, how, and when you deploy technology.

We will explore these and other questions in a series of articles on AI. We will separate fact from fiction and look at what’s possible today and what’s pie-in-the-sky. Most importantly, we’ll focus on the healthcare revenue cycle and get specific with use cases and benefits. We hope this series is inspirational and informative and will help you make good decisions for your people and business.

In this first article, we’ll define terms you’ve probably heard, give context to the topics surrounding AI, and lay a foundation for upcoming articles.

Definitions

Artificial intelligence (AI) is a field of computer science study that examines machine intelligence versus other forms of intelligence.1 It considers how machines might mimic human behaviors like physical movement, language, senses (seeing, hearing, etc.), and problem-solving. Many of us are exposed to AI every day without realizing it. We’ve grown used to AI-generated videos on YouTube and TikTok, chatbots capable of human-like interactions to help with complex tasks, and tools to generate content for us to write the perfect sentence.

Generative AI is artificial intelligence capable of generating text, images, videos, or other data using generative models, often in response to prompts. Generative AI models learn the patterns and structure of their input training data and then generate new data with similar characteristics.2 Examples of GenAI are ChatGPT’s text responses to questions or prompts and image creation from text prompts using Midjourney or Dall-e.

Machine learning (ML) is a branch of AI study and a fundamental component of AI systems. ML enables computers to imitate how humans learn by processing data and improving performance over time.4 For example, when you unlock your phone using your face or when Siri or Alexa responds to your voice, machine learning models trained to interpret and understand the shape of your face or the timbre of your voice enable those features. As ML-enabled systems acquire more data, models are updated for accuracy and better performance.

Large Language Model (LLM) is a machine learning model trained on massive amounts of textual data to make highly accurate predictions to generate content and perform a wide range of tasks.5
For example, medical researchers train large language models in healthcare on data from textbooks, research papers, and patient electronic health records to uncover patterns in disease and predict outcomes.6

LLMs can also:

  • Generate text
  • Translate between languages
  • Analyze sentiment from consumer feedback
  • Program robots to do physical tasks
  • Give a chatbot answers to human questions

LLMs also have downsides. Perhaps you heard of the lawyer who used ChatGPT to write a brief citing non-existent cases7 or ChatGPT’s problems with basic math. Model performance deviates over time with new data or when assumptions about data change. This deviation is called model drift or model decay, and it results in the model generating nonsensical responses to prompts previously handled correctly.

ChatGPT is the most famous large language model. Think of ChatGPT as a search engine on steroids rather than a higher form of intelligence. Thanks to massive computing power, it generates 100 billion words per day8 for humans in response to questions or prompts.

Data Mining is the process of searching and analyzing a large batch of raw data to identify patterns and extract useful information.9 Data is the lifeblood of AI, and the ability to see similarities, categories, or relationships in data means an AI system or model can be used for a specified purpose.

Thought starters

As stated earlier, there is a lot to process when it comes to AI. At a high level, the following are things to ponder on your journey with this much-hyped technology and its potential to change how we live and work. These thought starters will keep you grounded and offer context for discussions with coworkers and collaborators.

  1. Artificial intelligence is augmented intelligence.
    AI augments our abilities to analyze and make informed decisions or even move, see, or speak for us. It helps us make sense of a world drowning in data and allows us to use data to better humanity.
  2. When it comes to data, it’s garbage in, garbage out.
    Patterns and insights will only be usable if data is good and plentiful. The quality of data is paramount when evaluating any AI output. Bad data can introduce bias, incoherent responses, falsehoods, or harmful results.

    “We have to remember that AI can only be as useful as the data that informs it.” —Jon Clifton, Gallup CEO10

  3. Understanding intention will improve decision-making.
    Understanding the intention of those who build AI models and evaluating the data quality will enable you to take a more nuanced and informed stance when looking at results. When you see bias, for example, you can better interpret results with proper context.

    “Leaders have the potential to impose [their] will on a million decisions; that’s when it’s crucial to analyze precisely how data is used, and decisions are made.” —Cassie Kozyrkov, AI expert and former Chief Decision Scientist at Google11

  4. We won’t stop working anytime soon.
    Robots aren’t coming for our jobs. AI will make some jobs obsolete, but new jobs will emerge as AI expands and requires more human input and guidance to improve and become more valuable in more ways. Workers should consider AI an opportunity to learn new skills and grow professionally.
  5. Specialized AI will be more beneficial for business
    More companies will build models from proprietary data and find solutions to problems that have vexed them. Individual groups within organizations can develop their applications from company models that meet their specific needs, workflows, and requirements.

Final thoughts

AI is evolving quickly. Its practical uses are widespread and increasing daily, and its barriers are dropping. We are at the beginning of a significant shift in the way we create and communicate. Do you remember when email was the only form of digital communication in business? The pace of change over the last thirty years has been mind-boggling. What’s happening now and what will happen with AI will be even more transformative and awe-inspiring.

“The future is already here; it’s just not evenly distributed.”
-William Gibson

The next article in this series will examine how to get started with AI and why the revenue cycle is the best proofing ground in healthcare as we look at privacy, efficacy, and AI in medicine.
For more information about Janus solutions for automating the revenue cycle, drop us a line.

Read the rest of our AI blog series:

1 The New Healthcare C-Suite Agenda: 2024-2025 Sage Growth Partners, Jan 22, 2024

2 Wikipedia

3  Wikipedia

4 What is Machine Learning? IBM 

5 What are Large Language Models? IBM

6 What are Large Language Models? Nvidia

7  Here’s What Happens When Your Lawyer Uses ChatGPT, New York Times, May 27, 2023

8 Independent, February 12, 2024

9 What is Data Mining? Investopedia, February 23, 2024

10 Linkedin

11 Entrepreneur, March 1, 2024