The bigger challenge for education in a world of AI

Our education system values grades far more than learning. AI might finally be the technology that forces us to address this.

Chat GPT was a breakthrough event for Artificial Intelligence in the public sphere. For those who've followed AI for the last several years this may have been less of a shock, but to those not watching closely, the beta release of Chat GPT was probably the first time that the term "AI" had tangible meaning.

Since it's release, Chat GPT has "passed" several law school exams, and medical exams, and stirred up a huge amount of hyperbole about the future of AI. Reading these two sources gives a more grounded take on what was actually achieved by Chat GPT, and what help was given by humans, but this is just a marker of what the current technology is capable of.

Google, in their latest blog post on AI, state that:

Today, the scale of the largest AI computations is doubling every six months, far outpacing Moore’s Law

Twitter is awash with educators showing how Chat GPT has saved them hours of time when planning lessons; and educational publications are full of articles, such as this one, opining on what this means for the teaching and learning profession.

All of this is interesting but, for me, there's a more fundamental issue that we need to think about.

AI is a challenge to what it means to have created something. If I enter a prompt in to DALL·E 2 and produce some incredible artwork, have I created it? If I produce an essay for an exam using Chat GPT, is the grade mine? What if I only use it for research? What if I only use it to ask one question about the phrasing of one sentence that I wrote?

Trying to detect AI and treat it as we currently handle cheating will just create an arms race that will incorrectly penalise students for "AI assistance" when there was none, and fail to detect "AI assistance" where it was used to excess.

Students value their grades over the methods by which they achieve them. This is because we value the outcome (the grade) far more than the process (the learning, understanding, and skills that our students develop). There are plenty of teachers that buck this trend, and many care passionately about process, but in UK education, they are swimming against the tide.

For this tide to change, we need to rethink what our exams actually assess, and how we educate, what skills we actually value, and what knowledge is important. This is a huge challenge, and I don't have the hubris to pretend that I know what this looks like, but we do need to start thinking about it. It's probable that AI will create such uncertainty over what was and was not done by a student that the validity of the current exams system will be reduced to zero.

AI presents us an opportunity to reconsider how we educate, prioritising learning and understanding over exam scores. Make no mistake, this would be a fundamental shift from primary school, right through to university and into the workplace. Are we up to the challenge?