AI – not what you know but how fast you can learn

0
3

Source: Post Primary Teachers Association (PPTA)

“Currently, assessing outputs is a measure of learning. Learning needs time and effort. Outputs need time and effort. AI breaks that relationship so outputs are fast and easy. All of the training we’ve done in how to do research –  AI now does it for us and it is doing it well.”

Simon McCallum encouraged teachers to think about whether they were teaching classical musicians or DJs. “Is what we are teaching on the pathway to students’ careers or what those careers used to be?

Fluid intelligence prime objective

“The quality of an education is measured not by how much you know, but on how fast you can learn. We need to change education to focus on adaptability and learning, rather than memorisation and compliance. Fluid intelligence becomes our prime objective.”

He has observed students using AI and says there are three distinct groups. “There seems to be one group who are using it a lot to avoid learning. They are replacing the effort to learn with the effort to work out how to get AI to do the task they have been asked to do. There is another group who are not using AI because they have been told not to. Then there is another group who are using AI a lot and in interesting ways. They are not using it to replace themselves, they are using it to augment themselves.So when we assess them they are moving much much faster because they are building their learning on top of AI.”

AI coming for all activities

Simon McCallum says AI has been coming for a while but society is not ready for it. “Change is now constant and the rate of change is acclerating. If AI is a five times multiplier then only 20  percent of companies need to adopt it to change an industry.”

AI is not just language models – it is  coming for all our activities. Rabbit AI for instance trains a model to do actions for you. Alpha Geometry, developed by Google subsidary DeepMInd, can solve hard problems in Euclidean geometry. It achieved a silver medal in the International Mathematical Olympiad.

Hold close to human connections

Humanoid robots, developed by Tesla, were beginning to look more human, could take instructions in English, acted autonomously, and moved in a disturbingly human way.

“We are going to have to hold close what we value, such as strong human connections.”

Dr McCallum is currently employed by Victoria University of Wellington, the Norwegian University of Science and Technology, and Central Queensland University, winning teaching awards in both NZ and Norway. He has 25 years’ experience lecturing in Computer Science, AI and Computer Game Development.

MIL OSI

Previous articleHihi return to Shakespear Regional Park
Next articleCharter schools create unlevel playing field – a US teacher’s experience