In 2015, a nonprofit artificial intelligence lab was established in San Francisco. The founders of the new company wanted to protect against a future which they anticipated would be dominated by AI technology controlled by companies such as Google, Meta and Microsoft. Writing in the Washington Post, Nitasha Tiku states, ‘The nonprofit’s goal was to build AI software transparently and make its products open-source so the world could benefit’. OpenAI got off to a great start when various movers and shakers in Silicon Valley pledged over $1 billion dollars to launch its operation. Even Elon Musk got involved – that is before he grew bored and left the board in 2018 in order to focus on Tesla and SpaceX.
In November 2022, OpenAI launched ChatGPT and by January 2023, the software had over 100 million users worldwide. Using gargantuan amounts of data and incredibly powerful networks, the software is designed to mimic the human brain. OpenAI set out to build a system that comprehended human language in all of its complexities. ChatGPT is an extraordinarily powerful text generator that is capable of producing sophisticated prose. It is able to discuss the mystery of the incarnation or reimagine Beyonce’s lyrics in the form of a Shakespearean Sonnet. It can dash off computer code in an instance and debate the finer points of the Reformation with the conviction and authority of a learned Oxbriddge professor. Depending upon your viewpoint, it is either a diabolical Frankenstein-esque monster or an absolute Godsend for the procrastinating student facing yet another caffeine-fuelled all-nighter. Yes, it can even write a pretty convincing student essay but don’t tell the children!
Since the beginning of the year, ChatGPT has provoked a fierce debate in the pages of the Times Education Supplement and other learned journals. Some educationalists think ChatGPT constitutes the beginning of the end whilst others see it as a tool to be embraced by teachers and students alike. It has been suggested that it might even be able to evade anti-plagiarism software, though Turnitin.com argues, rather loftily, that it is in fact able to detect AI writing because it tends to be ‘extremely average’.
Critically, one should remember that students have found ways of cheating since time immemorial. Over a thousand years ago, cheating was said to be rife in the Chinese Civil Service examinations. Those who could not be bothered to learn endless Confucian aphorisms found ways to deceive their examiners despite the fact that such behaviour carried the death penalty. Examination boards and university faculties have long grappled with the problem of cheating. Notes sewn into the hem of a jacket, smart watches, earpieces, glasses and all manner of devices have been utilised in an attempt to gain that critical edge dishonestly. Online examinations have created new and extravagant ways of cheating. Indeed the dishonest student now has a whole toolbox of hi-tec tactics and gadgets at their disposal. No need to smuggle in cheat sheets or persuade your infinitely more intelligent identical twin to take your place. No need to make a suspiciously lengthy trip to the bathroom or glance over the shoulder of the class swot. A combination of AI and the move towards online examinations has made life so much easier…..or has it?
The reality is that AI text is great when it comes to generating factual content but it is pretty hopeless when it comes to sustaining a nuanced argument. It is not of this world and so will often provide explanations that are expressed with an awkwardness that should ring alarm bells for teachers who are familiar with the authentic voice of their students. AI induced text lacks tact, eloquence or grace. Its use of language is bombastic, amoral and based on a cold reductive logic that betrays its artificial nature. On occasion, it even serves up text that is racist or misogynistic. It has no soul and, in imitation, it fails to grasp what it is to be a sentient human being.
I am inclined to agree with those who think that the dangers of ChatGTP are probably somewhat overstated. It is up to schools and universities to propagate a culture of academic honesty. Furthermore, examinations which require children to harness higher order learning skills are unlikely to be undermined by cheating scoundrels intent on harnessing the power of AI! A focus on formative as opposed to summative assessment may provide an additional layer of protection against the high stakes ‘all or nothing’ scenario of the final examination.
Finally, perhaps it is time to rethink public examinations. The ability to withstand writer’s cramp and triumph in an angst-ridden race against the clock seems a curious way to measure our knowledge and understanding. Continuous assessment and the judgement of one’s teachers should count for something at least. Perhaps it is time to recreate Plato’s Symposium and place greater value on the ability to debate and develop ideas within an intellectually challenging context where rhetorical skills count for more than our ability to write.
The International Baccalaureate Organization (IBO) requires all member schools to have an Academic Honesty policy and so we have one here at Rossall. We talk to pupils about the importance of referencing sources and avoiding plagiarism/intellectual theft.
Academic integrity matters and those who try to cheat the checks and balances that are put in place by universities and schools rarely succeed. Just as it is never hard to spot the student who cuts and pastes chunks of text from Wikipedia or the BBC Bitesize website, I do not think that we will blind to the use of ChatGTP. Cheaters cheat themselves but institutions where cheating is endemic have failed to inculcate decent values in their student body. Ultimately, we are all responsible for the culture that exists within our schools – staff and pupils alike. A positive culture which encourages a culture of academic honesty is of greater value than a focus on staying one step ahead of technology.
This blog was written by AI or was it?
Headmaster of Rossall School