Tom, do you think the days of elearning are numbered, given that now we seem to have cracked the '2 sigma' problem with possibilities of AI powered conversational guidance and tutoring?
Great question, but I would say no in two senses, really:
1. AI-created learning experiences are e-learning; they are electronic learning. They are just different to what we think of as e-learning today. This is very similar to when everyone was treating video as not e-learning, despite the fact that it just is.
2. The AI tutors that I have seen fall astronomically short of human tutors. Just last week, I was very proudly being shown something from someone who had created a tool and brought it to market, only to have information that I absolutely knew was false presented to me.
Right now, AI is passable as a sort of slightly unreliable search engine, but I would never trust any information it gave me, making it useless as a tutor. Trust is a foundational requirement for effective tutor-based learning, and I honestly don't see the general populace trusting AI in the near future. As a side note, I hope this extends beyond the near future and we never fall into the world of attributing "trust" to any piece of software.
All of this said, I have absolutely no ability to predict the future, but when it comes to AI, month after month I just continue to adjust my expectations downward as things continue to be thoroughly unimpressive. Given how much of the two sigma problem material that is out there is driven by vendors, I am inherently suspicious of how accurate the suggestion is. I have to fall back on what I have seen until there is much more of a body of research around this.
Tom, do you think the days of elearning are numbered, given that now we seem to have cracked the '2 sigma' problem with possibilities of AI powered conversational guidance and tutoring?
Great question, but I would say no in two senses, really:
1. AI-created learning experiences are e-learning; they are electronic learning. They are just different to what we think of as e-learning today. This is very similar to when everyone was treating video as not e-learning, despite the fact that it just is.
2. The AI tutors that I have seen fall astronomically short of human tutors. Just last week, I was very proudly being shown something from someone who had created a tool and brought it to market, only to have information that I absolutely knew was false presented to me.
Right now, AI is passable as a sort of slightly unreliable search engine, but I would never trust any information it gave me, making it useless as a tutor. Trust is a foundational requirement for effective tutor-based learning, and I honestly don't see the general populace trusting AI in the near future. As a side note, I hope this extends beyond the near future and we never fall into the world of attributing "trust" to any piece of software.
All of this said, I have absolutely no ability to predict the future, but when it comes to AI, month after month I just continue to adjust my expectations downward as things continue to be thoroughly unimpressive. Given how much of the two sigma problem material that is out there is driven by vendors, I am inherently suspicious of how accurate the suggestion is. I have to fall back on what I have seen until there is much more of a body of research around this.
That's a blog post that could do with much amplification Tom. Get the basics right!