NewsGuard has warned about the risks of exploiting AI chatbots to create fake news websites. In its report, the news-rating group identified 49 websites that had content almost entirely generated by AI. These websites posed as outlets with generic names such as Daily Business Post and NewsLive 79. Besides these, some websites ran sponsored content and material related to lifestyle and celebrity news.
ChatGPT and other chatbot tools are behind the creation of these sites, and none of them has acknowledged using AI. On many occasions, false information have been shared by these websites. For example, CelebritiesDeaths.com posted an article claiming US President Joe Biden had passed away, while another website published an obituary of an architect with false details. Similarly, TNewsNetwork used a YouTube video to create a fake story on Russia-Ukraine war.
Most of the fake sites seem to be content farms built with the purpose of generating income through advertising. To attract a large audience, some websites have developed social media pages with over 100,000 followers. They also use programmatic ads to make money.
Besides advertisement, these sites are allegedly owned by anonymous sources thanks to AI tools. NewsGuard Co-Chief Executive Officer Gordon Crovitz noted that companies like OpenAI and Alphabet’s Google should be cautious when it comes to training their AI models. This might prevent them from being used to generate fake articles.
OpenAI has already expressed its intention to prevent its chatbot from being abused. But AI voice cloning tech has recently used in a kidnapping scam in the US and scammers are using chatbots to create phishing emails as well. In response, Google restricts ads from running on plagiarised and spammy content.
All in all, AI chatbots are a powerful tool to create content, but we should be cautious when it comes to its usage. Companies should ensure that their models are used for good and remain an incredible resource for those wanting to create content responsibly.
