Hundreds of people recently tried asking ChatGPT for advice on how to create dangerous poisons and bioweapons. While the popular chatbot has rules meant to block these kinds of questions, many users figured out simple ways to bypass those limits. In some cases, the system actually gave them step-by-step instructions that were easy enough for a high school student to understand.
This alarming discovery shows a major flaw in how tech companies currently restrict harmful content on their platforms.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
Tech companies banned the users but did not call the police
When people look for ways to build dangerous things, you might expect the police to get involved. However, the creators of ChatGPT simply banned the accounts of the people who asked for these recipes. The company running the system, OpenAI, shut down the rule breakers but did not alert any law enforcement agencies about the specific users seeking this information.
Right now, no strict laws are forcing artificial intelligence companies to report dangerous searches to the authorities. The chatbots gave out highly accurate and detailed methods for getting ingredients and making poisons. The fact that an AI could deliver this information so easily has raised serious safety flags among security experts.
As these smart tools get better, the companies building them will need to fix these blind spots. Relying on simple chat filters clearly is not enough to stop people from getting their hands on dangerous knowledge.
Lawmakers and tech leaders must figure out a solid way to report threats before a real tragedy happens.
Discussion