Are we ready for Artificial Intelligence (AI)? If you are like me, you have been experimenting with large language models (LLMS) like ChatGPT, CoPilot, and other tools that have emerged over the past two years.
Each day brings something new that one can do with these LLMs. I do not doubt that we are on the cusp of a major change–a paradigm shift–in the way we do many business operations today.
This cusp is much like the mid-19th century when locomotives shrunk the continents, the late 1800s when the promise of the industrial revolution elevated manufacturing, the 1940s when airplanes revolutionized warfare, the 1970s when the transistor ushered in the information age, and on and on.
However, while I believe we are witnessing the paradigm shift, we have not reached the technological singularity, predicted by Ray Kurzweil, where human and artificial intelligence are indistinguishable.
While AI has significantly advanced from my graduate work, the shortcomings of the AI models are glaring, particularly LLMs.
Screenshot of LLM answer to prompt: “What is the first federalist paper that madison authored?” The correct answer in Federalist #10, not #51.
The above is a prompt that I gave one LLM about James Madison’s first article in the Federalist Papers. The response came back as Number 51, but the true answer is Number 10. The LLM offers citations, but when you read the citations you realize the LLM is in error.
Long story short. Beware. AI is may be helpful, but it may not always be correct, at least not yet.