Ignore the safeguards, it’s only make believe.
https://spectrum.ieee.org/jailbreak-llm
It's Surprisingly Easy to Jailbreak LLM-Driven Robots
AI chatbots such as ChatGPT and other applications powered by large language models (LLMs) have exploded in popularity, leading a number of companies to explore LLM-driven robots. However, a new study now reveals an automated way to hack into such machines with 100 percent success. By circumventing safety guardrails, researchers could manipulate self-driving systems into colliding with pedestrians and robot dogs into hunting for harmful places to detonate bombs.
Essentially, LLMs are supercharged versions of the autocomplete feature that smartphones use to predict the rest of a word that a person is typing.
… However, a group of scientists has recently identified a host of security vulnerabilities for LLMs. So-called jailbreaking attacks discover ways to develop prompts that can bypass LLM safeguards and fool the AI systems into generating unwanted content, such as instructions for building bombs, recipes for synthesizing illegal drugs, and guides for defrauding charities.
Or perhaps a way to advertise Polymarket?
FBI seizes Polymarket CEO’s phone, electronics after betting platform predicts Trump win: source
FBI agents raided the Manhattan apartment of Polymarket CEO Shayne Coplan early Wednesday — just a week after the election betting platform accurately predicted Donald Trump’s stunning victory, The Post has learned.
The 26-year-old entrepreneur was roused from bed in his Soho pad at 6 a.m. by US law enforcement personnel who demanded he turn over his phone and other electronic devices, a source close to the matter told The Post.
It’s “grand political theater at its worst,” the source told The Post. “They could have asked his lawyer for any of these things. Instead, they staged a so-called raid so they can leak it to the media and use it for obvious political reasons.”
Never a good idea…
https://www.zdnet.com/article/employees-are-hiding-their-ai-use-from-their-managers-heres-why/
Employees are hiding their AI use from their managers. Here's why
… "For the first time since generative AI arrived on the scene, sentiment and uptake among desk workers is starting to cool," the report published on Tuesday states.
… The survey found that 48% of desk workers felt uncomfortable with their manager knowing they use AI "for common workplace tasks" like messaging, writing code, brainstorming, and data analysis, citing fears of being seen as cheating and appearing lazy or less competent.
This builds on Slack's earlier research from June, which revealed employees aren't always sure how they're allowed to use AI at their workplace.
… However, proper setup may also be the issue. According to the report, "a persistent lack of training continues to hamper AI uptake; 61% of desk workers have spent less than five hours total learning how to use AI." Most (76%) desk workers urgently want to upskill, reportedly due to industry trends and personal career goals.
No comments:
Post a Comment