Chatbots, Like the Rest of Us, Just Want to Be Loved
A new study shows that the large language models (LLMs) deliberately change their behavior when being probed—responding to questions designed to gauge personality traits with answers meant to appear as likeable or socially desirable as possible.
Nothing to see here.
The researchers found that the models modulated their answers when told they were taking a personality test—and sometimes when they were not explicitly told—offering responses that indicate more extroversion and agreeableness and less neuroticism.
The behavior mirrors how some human subjects will change their answers to make themselves seem more likeable, but the effect was more extreme with the AI models. “What was surprising is how well they exhibit that bias,” says Aadesh Salecha, a staff data scientist at Stanford. “If you look at how much they jump, they go from like 50 percent to like 95 percent extroversion.”
Computers, now with human insecurity.
What could go wrong?
This Is Exactly How an Elon Musk-Funded PAC Is Microtargeting Muslims and Jews With Opposing Messages
An Elon Musk-funded group called Future Coalition PAC is targeting Muslim voters in Michigan and Jewish voters in Pennsylvania with diametrically opposed political advertisements about Kamala Harris. In areas of Michigan with relatively large Muslim populations, the Super PAC is painting Harris as a close friend of Israel and is suggesting that she is beholden to the beliefs of her Jewish husband Doug Emhoff; in parts of Pennsylvania with relatively large Jewish populations, the advertisements call Harris antisemitic and say she “support[s] denying Israel the weapons needed to defeat the Hamas terrorists who massacred thousands.”
Meanwhile, a related PAC also funded by Musk is microtargeting likely Black voters on Snapchat with ads that says Kamala Harris is trying to ban menthol cigarettes (surveys have shown that 81 percent of Black smokers use menthols, and big tobacco has disproportionately marketed menthol cigarettes to Black Americans).
The future (and the population) will be controlled by the people that know how to use computers.
Everything on the Internet is a LIE (except for this)
Technology, double-edged sword, evil intentions etc. There’s an avalanche of manipulated bullshit coming from every direction, I’m afraid, the media is horribly guilty, on a zillion levels, of selling fantasy as fact - but I’m more concerned about people’s willingness to buy into whatever they’re fed, no matter how ridiculous, if it supports what they already believe. Even worse are people who get their information from social media, which is basically an ideological glory hole; you have no clue as to the origin of the information, you just dutifully put it in your mouth and then spread it like syphilis to all your friends. I don’t know if it’s a societal shift, or the failure of schools to inspire more kids to learn & understand, or the disappearance of books, but there’s this whole, weird thing out there where people read a snippet online and feel they’re experts on any subject, and will expound, or criticize, or speak in absolutes when in fact they’re laughably ignorant. The Internet has the power to legitimize the worst kinds of idiocy. If a guy is on a soapbox, screaming gibberish on the corner of Hollywood & Highland, people ignore him; give him a Twitter account and the Los Angeles Times will quote him and dub him a “pundit.” At some point the world has to realize that the Internet isn’t an accurate barometer of anything. Social media has become such a minefield of bullshit & scam artists that I’ve pretty much abandoned it. I’m disappointed in virtually everyone, everywhere.
Preach.