A new artificial intelligence chatbot able to produce generative text and even college level essays could open the floodgates for propaganda and disinformation campaigns.
OpenAI’s ChatGPT is an AI chatbot that was launched on December 1 and exceeded one million users in only five days.
However, a report contributed to by OpenAI policy researcher Girish Sastry and former policy researcher Matthew Gentzel, revealed just how dangerous generative language models like ChatGPT could be in the wrong hands.
The report explored “potential misuses of language models for disinformation campaigns” and outlined how disinformation could be spurred by propagandists.
Those taking advantage of the AI chatbot system would be able to capitalise on its fast delivery of text and content to spread more personalised propaganda at a quicker and cheaper rate.
Stream more science news live & on demand with Flash. 25+ news channels in 1 place. New to Flash? Try 1 month free. Offer ends 31 October, 2023
Potential real-world challenges from generative AI texts include the emergence of a “larger number and more diverse group of propagandists ” as well as concerns that messages could grow more credible meaning “propaganda is less discoverable”, according to findings of the report.
In the wake of growing concerns, 22-year-old university student Edward Tian developed an app, GPTZero, to detect when an essay or text was built using the AI bot.
Mr Tian said his tool uses perplexity – “randomness of a text to a model” – as one of the metrics to test against AI. Lower levels of randomness detect AI influence.
“Hopefully will be adding a few more tests and factors to improve the model,” Mr Tian tweeted on January 4.
Despite reservations from some, others believed Australia needed to adopt the technology.
University of Technology Sydney professor Leslie Loble said the nation needed to pay attention to AI.
Ms Loble suggested in a report that Australia was “flying blind” when it came to knowing how AI and edtech (educational technology) applications work and how to operate them safely.
Meanwhile, University of Sydney’s Marcel Scharth said ChatGPT was a big “leap in technology”.
“For those of you at home that are worried about this technology, let me say I encourage you to be curious, explore it and try it for yourself. It could be useful to you today so use it to your advantage,” Mr Scharth told Sky News Australia’s Andrew Bolt in December.
Leave a Reply