AI means so many different things to so many people, right? It’s a bit like trying to describe a rainbow – everyone sees something a little different. We need better ways to understand and talk about it. Luckily, a gifted geek and cultural anthropologist, Drew Breunig, has come up with a neat categorization of AI technology into three insightful use cases: gods, interns, and cogs. Let’s dive into this fascinating framework for understanding AI’s evolving role in our lives.
AI Use Cases: A New Framework
AI Gods: The Super-Intelligent Entities
Imagine super-intelligent, artificial entities that operate autonomously. This is the “artificial general intelligence” (AGI) that some researchers are striving to build, even as they raise concerns about the potential existential threat it poses to humanity. These AI gods are essentially human replacement use cases, demanding massive computational resources, enormous amounts of energy, and raising serious concerns about environmental impact.
AI Interns: Supervised Collaborators
These are the supervised co-pilots collaborating with experts, focusing on the grunt work. Think ChatGPT, Claude, Llama, and similar large language models (LLMs). They’re designed to be used and supervised by experts. They have a high tolerance for errors because the humans overseeing them catch and correct potential mistakes. Interns handle the tedious tasks: remembering documentation, navigating references, filling in details after the bigger picture is defined, and even assisting with idea generation. They’re the prevalent AI technology in use today, and we’re gaining insights into how these human-machine partnerships work in the real world.
AI Cogs: The Specialized Machines
These are the lowly machines optimized for a single task. They often work as part of a pipeline or interface, efficiently performing a specific function without much broader understanding or cognitive abilities. Think of them as highly specialized, dedicated workers.
AI in Healthcare: An Example
One exciting area for AI application is healthcare. For instance, AI algorithms can analyze retinal scans exceptionally quickly to identify symptoms requiring immediate treatment. But what about the diagnostic process itself?
The Surprising Study on AI Diagnosis
A recent study published in the Journal of the American Medical Association provided interesting results. Researchers investigated whether ChatGPT could boost the diagnostic skills of practicing physicians. The conclusion? While ChatGPT alone performed better than the physicians, the presence of conventional resources didn’t significantly improve the physicians’ diagnostic accuracy. This highlights the importance of “prompt engineering” – understanding how to effectively interact with LLMs to get the best results. Doctors sometimes clung to their initial diagnoses even when ChatGPT offered better alternatives, and some weren’t fully harnessing the tool’s potential.
The Impact on Job Satisfaction
Another intriguing discovery was the effect of AI collaboration on the job satisfaction of material scientists. Using AI to assist in research led to a notable increase in the discovery of new materials and patent filings, with the AI performing a significant portion of the initial idea generation. However, this increased productivity seemed to come at the cost of reduced job satisfaction. This demonstrates a crucial point: AI’s role in augmenting our capabilities isn’t always a straightforward win.
This all points to a crucial need for a better understanding of AI use cases. The next time you encounter AI, consider whether it fits into the category of god, intern, or cog. This categorization can offer helpful insight into how the technology is meant to be used and what impact it’s likely to have.
What are your thoughts on these AI use cases? Leave a comment below and share this article with your friends!
“`
