Major corporations like Microsoft and Alphabet are in the race to use generative AI capabilities to improve their chatbots and search engines. This new AI is being trained on data obtained from all over the Internet, giving users a one-stop-shop for their queries. However, a recent report from Israeli venture capital firm Team8 has warned that corporate secrets, confidential information, and even user data could easily fall into the wrong hands if these tools are not safeguarded and security measures are not in place.
ChatGPT, a generative artificial intelligence tool, is at the risk of exposing confidential data if it falls into the wrong hands. According to the Team8 report, not only could hackers use ChatGPT to access sensitive corporate information, exploiting the chatbot’s capabilities could also potentially lead to data leaks, lawsuits and other adverse consequences. The report also pointed out that training models using data obtained from the Internet makes it difficult to erase any sensitive material.
The Team8 report was endorsed by Michael Rogers, the former head of the US National Security Agency and US Cyber Command, and also included input from Microsoft’s Corporate Vice President Ann Johnson. Microsoft has invested billions in Open AI, an AI platform which developed ChatGPT, and is one of the powerful corporations that is leveraging on new AI technologies.
However, the report noted that chatbot queries aren’t being fed into large-language models to train AI and debunked several previous reports to the effect. Despite this, the fear of third party applications exploiting generative AI to access private data remains a worrying reality. Information shared through third-party applications still leave companies open to the risk of the information being exploited by hackers.
The Team8 report has identified three high-risk issues related to implementing generative AI tools, and these need to be dealt with carefully. Companies using AI platforms like ChatGPT also need to stay alert and ensure that the required safeguards are in place to protect confidential customer data and trade secrets.
