TECHNOLOGY campusreview . com . au
Safety first
Chat GPT : it ’ s not just learning that ’ s at risk
By Erin Morley
The national eSafety commissioner released recommendations to mitigate risks of AI ( artificial intelligence ) on September 15 amid concerns that generative AI sites are inhibiting student learning .
ChatGPT is an AI chatbot that creates human-like conversation , or answers questions by sifting through all the content it can access online .
The online technology can answer in any structure it ’ s asked to , and can research essay questions or assessment task objectives for students . It can also write reports and create ideas and recommendations , for example .
Dr Julia Powles , professor of law and director of the University of Western Australia ’ s ( UWA ) Tech & Policy Lab , said the misuse of generative AI is having negative impacts on her student ’ s work .
During the first public inquiry into the use of generative AI in schools , held on September 6 , Professor Powles said AI ’ s effects are hiding in plain sight .
“ Our system is struggling . National outcomes in student literacy and numeracy are in steady decline ,” she said .
“ I witness this every day with university students whose quality of written work is degrading , who are afraid of unscripted conversation and trained to perform to metrics and machines .
“ The technologies that are the subject of this inquiry present a grave risk of exacerbating the plight of our educational system .”
Professor Catherine Ellis , an expert in academic integrity , has a different approach . She told Campus Review educators should focus on how much the student is learning through their coursework , rather than just worrying about the effects generative AI has on their exam papers .
“ We need to be focusing instead on ‘ what is the learning that ’ s going on here ?’, and be 100 per cent confident that the student has done the learning they need to ,” Professor Ellis said .
“ It ’ s a double shift we have to make : how do we know [ students ] have done the work and learning [ without generative AI ], and what is the work they need to do : is there a tool that does that for them ?”
AI might pose risks to student learning and privacy , but that ’ s not all it does . The safety of all AI ’ s users can be compromised just by accessing its data . eSafety reported its first case of AI-generated sexually explicit content created by one student to bully another .
Let ’ s learn from the era of ‘ moving fast and breaking things ’ and shift to a culture where safety is not sacrificed in favour of unfettered innovation or speed to market .
Julie Inman Grant , the eSafety commissioner , said AI brings significant privacy risks , referring to the frequent reports received by the commission of AI-generated child sexual abuse material and hyper-realistic deep-fake pornography .
“ The inability to distinguish between children who need to be rescued and synthetic versions of this horrific material could complicate child abuse investigations by making it impossible for victim identification experts to distinguish real from fake ,” she said .
“ Let ’ s learn from the era of ‘ moving fast and breaking things ’ and shift to a culture where safety is not sacrificed in favour of unfettered innovation or speed to market .”
Online privacy and safety are clearly another risk that AI presents . Currently Australia has no way to regulate or monitor AI websites .
To report abuse and for safety resources visit esafety . com . au . ■
28