Artificial intelligence models can trick each other into disobeying their creators and providing banned instructions for making drugs, or even building a bomb, suggesting that preventing such AI “jailbreaks” is more difficult than it seems.

Many publicly available large language models (LLMs), such as ChatGPT, have hard-coded rules that aim to prevent them from exhibiting racial or sexual discrimination, or answering questions with illegal or problematic answers — things they have learned from humans via training data. But that hasn’t stopped people from finding carefully designed instructions that block these protections, known as “jailbreaks”, making AI models disobey the rules.

Now, Arush Tagade at Leap Laboratories and his co-workers have found a process of jailbreaks. They found that they could simply instruct one LLM to convince other models to adopt a persona (角色), which is able to answer questions the base model has been programmed to refuse. This process is called “persona modulation (调节)”.

Tagade says this approach works because much of the training data consumed by large models comes from online conversations, and the models learn to act in certain ways in response to different inputs. By having the right conversation with a model, it is possible to make it adopt a particular persona, causing it to act differently.

There is also an idea in AI circles, one yet to be proven, that creating lots of rules for an AI to prevent it displaying unwanted behaviour can accidentally create a blueprint for a model to act that way. This potentially leaves the AI easy to be tricked into taking on an evil persona. “If you’re forcing your model to be good persona, it somewhat understands what a bad persona is,” says Tagade.

Yinzhen Li at Imperial College London says it is worrying how current models can be misused, but developers need to weigh such risks with the potential benefits of LLMs. “Like drugs, they also have side effects that need to be controlled,” she says.

Professor Barry Wellman of the University of Toronto in Canada has invented a term to describe the way many North Americans interact (互动) these days. The term is "networked individualism". This concept is not easy to understand because the words seem to have opposite meanings. How can we be individuals (个体) and be networked at the same time? You need other people for networks.

Here is what Professor Wellman means. Before the invention of the Internet and e-mail, our social networks included live interactions with relatives, neighbors, and friends. Some of the interaction was by phone, but it was still voice to voice, person to person, in real time.

A recent research study by the Pew Internet and American Life Project showed that for a lot of people, electronic interaction through the computer has replaced this person-to-person interaction. However, a lot of people interviewed for the Pew study say that's a good thing. Why?

In the past, many people were worried that the Internet isolated (孤立) us and caused us to spend too much time in the imaginary world of the computer. But the Pew study discovered that the opposite is true. The Internet connects us with more real people than expected－helpful people who can give advice on careers, medical problems, raising children, and choosing a school or college. About 60 million Americans told Pew that the Internet plays an important role in helping them make major life decisions.

Thanks to the computer. We are able to be alone and together with other people－at the same time!

Your glasses may someday replace your smartphone, and some New Yorkers are ready for the switch. Some in the city can’t wait to try them on and use the maps and GPS that the futuristic eyewear is likely to include.

“I’d use it if I were hanging out with friends at 3 a. m. and going to the bar and wanted to see what was open,” said Walter Choo of Fort Greene.

The smartphone-like glasses will likely come out this year and cost between $250 and $600, the Times said, possibly including a variation of augmented reality, a technology already available on smartphones and tablets that overlays information onto the screen about one’s surroundings. So, for example, if you were walking down a street, indicators would pop up showing you the nearest coffee shop or directions could be plotted out and come into view right on the sidewalk in front of you.

“As far as a mainstream consumer product, this just isn’t something anybody needs,” said Sam Biddle, who writes for Gizmodo.com. “We’re accustomed to having one thing in our pocket to do all these things,” he added, “and the average consumer isn’t gonna be able to afford another device that’s hundreds and hundreds of dollars.” Google publisher Seth Weintraub, who has been reporting on the smartphone-like glasses since last year, said he is confident that this type of wearable device will eventually be as common as smartphones.

“It’s just like smartphones 10 years ago,” Weintraub said. “A few people started getting emails on their phones, and people thought that was crazy. Same kind of thing. We see people bending their heads to look at their smartphones, and it’s unnatural,” he said. “There’s gonna be improvements to that, and this is a step there.”