Scores of countries worldwide have banned artificial intelligence-driven ChatGPT amid privacy concerns, with Italy among the first to ban the platform temporarily.
Italy’s ban on ChatGPT has prompted other European nations and the world to scrutinise AI tools and enforce regulations around their use, which begs the question if or when South Africa could employ similar regulations.
South African futurist Charlotte Kemp said that before the country could consider banning AI-based platforms like ChatGPT, it would have to consider why other countries were banning it.
“Open AI has accessed only public information, so there shouldn't be any of those kinds of data privacy issues. One of the other comments that get thrown around a little bit is about the misinformation and countries, like… like Russia and China, which are blocking it over a fear of misinformation.
Kemp said Kemp said deep-learning models that could access all the data were not creating misinformation, as the information or content would have been created by a human being. All ChatGPT was doing was accessing the information faster than a human would.
.She added that If South Africa had to follow other nations by introducing a temporary ban on AI-based platforms until better regulation around the technology was in place, it woudn’t help much as it was already accessible to people everywhere.
“Even if a whole bunch of countries decided to pause for a while, there are going to be other countries still streaming head in terms of the development of the later versions of ChatGPT, but it’s already making a huge impact in the world, but we need to find better and safer ways to use it,” Kemp said.
Pal Aaserudseter, the engineer for cybersecurity firm Check Point, explained that Italy’s temporary ban was related to the General Data Protection Regulation because of a ChatGPT data leak on March 20 and that ChatGPT did not age-restrict usage.
“To specify, the data leak happened because of a technical error that exposed chat topic history to other users, but also exposed user billing and partial credit card information of users. In addition, Italy and other nations are concerned about uploaded user data possibly being stored in ChatGPT and that others can access their data,” Aaserudseter said.
Despite the leak, ChatGPT explicitly states that user data uploaded to the platform is only in memory and deleted after use, unless the user explicitly states the data should be saved.
Aaserudseter said the potential for data misuse was by far one of the greatest concerns regarding ChatGPT.
“We need regulations in place to make sure that OpenAI, Microsoft, Google and others developing AI technologies make sure that access, usage, privacy and data are protected and controlled in a secure fashion. Right now, it’s a loaded gun that anyone can use for whatever purpose they serve,” Aaserudseter said.