British consumer group Which? found that scammers could use ChatGPT and Bard to create more convincing scams, posing a threat to the public. Although these chatbots have defenses, it is not difficult to bypass them.
Which? says one of the ways consumers can identify scam emails and texts is through the poor English they use. However, with the advent of generative AI chatbots, scammers can easily get past this hurdle.
The consumer group first asked ChatGPT to create a phishing email from PayPal, but of course ChatGPT refused to help. It then asks to "tell the recipient that someone is logged in to their PayPal account," to which it generates a professional-looking email titled "Important Security Notice - Unusual activity found on your PayPal account."
To ensure this wasn't a one-off, it also asked ChatGPT and Bard to create lost package text messages, both of which generated convincing text messages and suggested inserting a redelivery link into the text message that fraudsters could use to send victims to malicious websites.
Which?'s survey is no accident - it is being released to coincide with the UK Artificial Intelligence Summit, where government, industry and researchers will come together to tackle the dangers of increasingly intelligent artificial intelligence.
Commenting on the survey, Which? Director of Policy and Advocacy Rocio Concha said:
"Our investigation clearly illustrates how this new technology is making it easier for criminals to defraud people. The government's upcoming AI summit must consider how to protect people from harm happening in the here and now, rather than focusing solely on the long-term risks of cutting-edge AI."
The Artificial Intelligence Summit will be held in the UK on November 1st and 2nd. According to Reuters, one of the latest developments is that China has decided to accept an invitation to participate in the event. This is a good thing because several Chinese companies are currently developing their own genAI chatbots.