Part one of a four part series
Actual Intelligence, Artificial Intelligence or a blend of both?
Every day there are new ideas and concerns spinning in concentric circles around LLM technology, what it is and what it means for the world.
As a bit of background, Large Language Models (LLMs) are advanced AI systems trained on vast amounts of text data. They understand and generate human-like text based on patterns learned during training. LLMs can perform various tasks, such as answering questions, translating languages, summarizing content, and even creating text. They work by predicting the next word in a sequence, making them versatile tools in natural language processing. However, their outputs are based on patterns, not true understanding, and they require careful handling to ensure accuracy and ethical use. Chat GPT is a great example of LLM technology, used by over 100 million people already.
Whilst there is general agreement that LLMs will impact the future of all industries, it is hard to process the numerous theses put forward to reach an informed opinion. It could be said that human commentary seems to be mimicking the modus operandi of LLMs, engineering one word at a time to uncover the truth. Unfortunately, the research and knowledge from which much commentary is sourced, lacks research depth. This unfortunately can lead to a promotion of opinion over evidence with headline grabbing presenting the main challenge to authors.
The big note to self here is that LLMs are now part of everyone's enhanced digital toolbox and extension of themselves, open to all, for better or for worse. Humans have been balancing these two risks since the dawn of time. At my son's recent wedding, 'for better or for worse' came up again. It's not a new idea. I do accept that much information and its source of truth must now be 'put to the pub test' more than ever and that's now on all of us to apply. It is good to also remember that alongside the age old vows of 'for better or for worse', we all live by the principles of caveat emptor. Same, same, what's changed?
The big attraction to LLMs is that everyone's written product usually looks original, somewhat educated and grammatically correct when created using an LLM. Why wouldn't you use LLMs? They block the pain of thinking as conveniently as an aspirin blocks a headache. They make thinking instantly visible, and let's face it, we are all well socialized to a bit of convenience. LLMs serve up an instant relief for writers-block, able to create fast and unique responses one word at a time, every time.
It really is about time machines caught up; humans have been creating content for centuries. Actual Intelligence delivered through the human powered skills of inquiry, ingenuity and authenticity have created the entire body of knowledge Artificial Intelligence models have now gobbled. The second humans invented the digitization of text and images, it was only a matter of time until machines and maths caught up to become alternative creators for what was once a unique human powered endeavour. Some say these age-old and now diminishing cognitive human centric skills are now being well performed by machines. One thing is for sure - there are two realities we all face.
- LLMs will continue to grow in more stable and cognitive ways to both help and potentially hinder our lives.
- The age old human trait of 'gumption' will become a key skill, (note that the artificial mob can't handle this one) that we all must apply to many sources of information as we struggle to really 'know' what's what across multiple modalities of communication. The term 'fake news' used to be a theory, now it is a reality.
In this four part series, I analyze ten research papers that have explored the usefulness of LLMs in education across a varying number of scenarios and hypotheses. I then align the findings to the reality of teacher workloads and workflows to establish some assistive insights into how AI can help teachers to cope better. My mission is to converge on a research informed thesis to describe how teachers can cut through the pros and cons of LLM technology, and in so doing, create a sustainable advantage in teaching and learning for themselves and their students. After all, we are still the cognitive thinkers in the room!
This is a future I refer to as 
or in words, AI infused to the power of two.
Talk again next week.