Image courtesy of Barbara Zandoval on Unsplash
Learning about ChatGPT is important. It is an AI chatbot that answers questions that we do not have answers to; therefore trust in it is critical. Security researcher, Johann Rehberger recenty made an important discovery about ChatGPT.
Hopefully we all know that AI stands for Artificial Intelligence and it is less fact-based than most would think. The intelligence is gathered from the internet and is continually updated and includes a lot of misinformation, what used to be called false statement or lies.
Rehberger discovered that the memories that ChatGPT can be tricked into creating false memories. Since the internet is the playground of malicious attackers, this is something they would have access to. While OpenAI has provided fixes, it has yet to produce preventative measures to remove this problem from happening at all.