August 16, 2023
November 2022 brought another AI disruption on the heels of the last one: ChatGPT, a generative artificial intelligence (GAI) bot that can often out-perform students on common assignments and get 5s on most of the AP exams, immediately became a global sensation, garnering 100 million unique visitors in 2 months. This was just the beginning of a disruption that will not end: GAI is here to stay and its abilities will continue to accelerate, dramatically impacting education and changing the world forever.
The abilities of these bots, which now include representatives from Google, Anthropic, Inflection, and Meta, to produce more content that is becoming more factually correct and authored in the voice of the user will grow. And so will the strength of companies that build on top of their technology to excel at specific tasks, such as writing student papers undetected.
They will also strengthen these bots to reason, perceive situations, solve problems, plan responses, and apply their learning to new situations. These are all common characteristics of human intelligence, and students have already started using these bots as tutors.
Today, we largely think of GAI as text-to-text (the machines writing the papers), but these tools are also starting to generate incredible images, short videos, and 3-D environments. Bot inputs and outputs will become multimodal: combinations of text, voice, image, and video will be used in the input as well as the output. When this occurs seamlessly, full spatial environments that are uniquely suited to each individual’s interests and desires will be easily produced and refined as users interact with them, creating incredible dynamic and interactive environments. We can even see the rudiments of one created by an 11-year-old.
The production of dynamic multimodal environments will be accompanied by the development of bots that progress more and more toward human-level intelligence. They will offer opportunities for machines to perceive situations (such as learner’s needs), respond (develop a plan to meet those needs), and respond again (meet the needs), continually perceiving the individual’s changed situation, iteratively developing responses, and adjusting to meet the new needs of the individual learner. As this happens and the rate of “hallucination,” which refers to the bots stating incorrect facts, declines, we will have fully intelligent tutoring systems (ITS) that will support individualized student…
