If you described your symptoms to me as a business leader and I typed them into ChatGPT, would you want me to generate and prescribe a treatment plan for you, sending orders to your local pharmacist — without consulting a doctor? What if you were offered a trade: The top data scientists in the world will join your organization, but with the catch that every one of your business experts must join your competitor, leaving only data to work with and no experts to provide context?
The Role of Language Models and Expertise
In the era of AI, the public square is filled with voices touting the opportunities, risks, threats, and recommended practices for adopting generative AI — especially language models such as GPT-4 or Bard. New open-sourced models, research breakthroughs, and product launches are announced daily. However, the emphasis on the capabilities of language models neglects the importance of expertise in making them truly useful.
“For language models, this goes a step further and can be misleading, because models cannot only recite related words, but underlying documents, frameworks, phrases, and recommendations that have been written by experts.”
Language models can generate outputs based on correlations between given inputs and previously seen patterns. For example, they can create a new recipe based on correlations between ingredients and descriptions. However, they lack true knowledge and understanding of what tastes good. The generated recipe may be statistically likely, but it is ultimately thanks to the expertise of real chefs whose recipes were part of the model’s training data.
“Language models are powerful, and the secret ingredient to making them useful is expertise.”
The phrase “correlation does not equal causation” is well-known in the field of data analysis. While machines excel at identifying correlations and patterns, human expertise is necessary to determine causations and inform decision-making. Language is only the first step in the human learning process. We gain knowledge and understanding through language, but it is the expertise that allows us to apply that knowledge effectively.
“The most common misconception in discussions around AI and ML is that data is the most critical element — but expertise is the most critical element. Otherwise, what is the model learning?”
Distilling Expertise into Machines
Machine learning (ML) and machine teaching are sub-disciplines focused on translating human expertise into machine language. ML aims to equip machines with the ability to learn, while machine teaching focuses on equipping humans to teach machines effectively.
“The process of building AI solutions should begin with the question of what expertise is most important to the organization.”
Expertise is the foundation from which AI systems should be developed. Data alone is not sufficient. It is the expert-driven decision-making that guides the development of AI models and algorithms. By translating expertise into machine language, autonomous systems can be created, freeing up human capacities for more nuanced decision-making and discovery.
Organizations must assess the level of risk associated with losing expert knowledge and identify areas where offloading expert-driven decisions to machines can lead to a significant upside. Evaluating which aspects of expertise can be translated into machine language is crucial in this process.
“By the year 2030, household names we now revere will have joined the ranks alongside Blockbuster because they chose to fast follow…”
Looking towards the future, organizations must not wait to react to AI advancements. Instead, leaders should envision the market potential that would require others to scramble to catch up. Investing in transferring operationalized expertise to machines and setting a bold vision for the future will position organizations for success in the era of autonomous transformation.
