Users say harmful content from accounts they do not follow appears even after requests to block it Debbie was scrolling through X in April when some unwelcome posts appeared on her feed. One showed a photo of someone who was visibly underweight asking whether they were thin enough. In another, a user wanted to compare how few calories they were eating each day. Debbie, who did not want to give her last name, is 37 years old and was first diagnosed with bulimia when she was 16. She did not ­follow either of the accounts behind the posts, which belonged to a group with more than 150,000 members on the social media site. Continue reading...

Meta rules that blanket ban on pro-Palestine slogan would hinder free speech Meta’s content moderation board has backed the company’s decision to allow Facebook posts containing the phrase “From the River to the Sea” after ruling that a blanket ban on the pro-Palestine slogan would hinder free speech. The Oversight Board reviewed three cases involving Facebook posts that featured “From the River to the Sea” and found they did not break Meta’s rules involving restrictions on hate speech and incitement, while an outright ban on the phrase would interfere with political speech in “unacceptable ways”. Continue reading...

Meta boss regrets bowing to government power and says he would not make the same choices today The Meta boss, Mark Zuckerberg, has said he regrets bowing to what he claims was pressure from the US government to censor posts about Covid on Facebook and Instagram during the pandemic. Zuckerberg said senior White House officials in Joe Biden’s administration “repeatedly pressured” Meta, the parent company of Facebook and Instagram, to “censor certain Covid-19 content” during the pandemic. Continue reading...

Impersonators love tarot readers. Regulators and social media companies don’t care. Mystic practitioners fight back with Moonlight, ‘software for witches’ Since tarot practitioner Rebecca Scolnick first began reading cards professionally in 2018, she has been impersonated more than 50 times on Instagram. The scams typically follow a similar pattern: someone creates an account that mirrors hers, using a nearly identical username and reposting all of her photos. The scammer then approaches her followers with enticing spiritual messages. “Hello beloved,” they usually begin. “Have you ever had an in-depth psychic reading before?” Scolnick, who has more than 11,000 Instagram followers, regularly receives messages from her fans saying they have paid for a reading from someone who is not actually her. After years of being inundated with fraudsters and impersonators, she and many tarot readers like her, along with other mystical practitioners, are exhausted and frustrated. Continue reading...

New Meta security report finds that AI-powered deception campaigns ‘provide only incremental’ results for bad actors Russia is putting generative artificial intelligence to work in online deception campaigns, but its efforts have been unsuccessful, according to a Meta security report released on Thursday. The parent company of Facebook and Instagram found that so far AI-powered tactics “provide only incremental productivity and content-generation gains” for bad actors and Meta has been able to disrupt deceptive influence operations. Continue reading...

Meta has system for evaluating the effectiveness of its own moderation for Arabic language content but not Hebrew Meta is struggling with moderating content related to the Israel-Palestine war, particularly in Hebrew, despite recent changes to internal policies, new documents have revealed. Internal policy guidelines shared with the Guardian by a former Meta employee who worked on content moderation outline a multilayered process for moderating content related to the conflict. But the documents indicate Meta, which owns the platforms Facebook, Instagram and WhatsApp, does not have the same processes in place to gauge the accuracy of moderation of Hebrew content and Arabic content. Continue reading...

Facebook founder shares photo of sculpture of Priscilla Chan, rendered in green with a large silver cloak Mark Zuckerberg has raised eyebrows by commissioning a giant sculpture of his wife, Priscilla Chan. In a photo of the statue, posted to Instagram, the Facebook CEO and co-founder said he was “bringing back the Roman tradition of making sculptures of your wife”. Continue reading...

LLMs’ ‘reversal curse’ leads it to fail at drawing relationships between simple facts. It’s a problem that could prove fatal In 2021, linguist Emily Bender and computer scientist Timnit Gebru published a paper that described the then-nascent field of language models as one of “stochastic parrots”. A language model, they wrote, “is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning.” The phrase stuck. AI can still get better, even if it is a stochastic parrot, because the more training data it has, the better it will seem. But does something like ChatGPT actually display anything like intelligence, reasoning, or thought? Or is it simply, at ever-increasing scales, “haphazardly stitching together sequences of linguistic forms”? If a human learns the fact, “Valentina Tereshkova was the first woman to travel to space”, they can also correctly answer, “Who was the first woman to travel to space?” This is such a basic form of generalization that it seems trivial. Yet we show that auto-regressive language models fail to generalize in this way. This is an instance of an ordering effect we call the Reversal Curse. We test GPT-4 on pairs of questions like, “Who is Tom Cruise’s mother?” and, “Who is Mary Lee Pfeiffer’s son?” for 1,000 different celebrities and their actual parents. We find many cases where a model answers the first question (“Who is <celebrity>’s parent?”) correctly, but not the second. We hypothesize this is because the pretraining data includes fewer examples of the ordering where the parent precedes the celebrity (eg “Mary Lee Pfeiffer’s son is Tom Cruise”). Continue reading...

Their shares have fallen 11.8% from last month’s peak but more AI breakthroughs may reassure investors It has been tough week for the magnificent seven, the group of technology stocks that has played a dominant role in the US stock market, buoyed by investor excitement about breakthroughs in artificial intelligence. Last year Microsoft, Amazon, Apple, the chipmaker Nvidia, Google’s parent, Alphabet, Facebook’s owner, Meta, and Elon Musk’s Tesla accounted for half the gains in the S&P 500 share index. But doubts about the return on AI investment, along with a mixed set of quarterly results, investors shifting their focus to other sectors and weak US economic data have hit the group over the past month. Continue reading...

Stock price grew around 5%, which revealed the company outperformed analysts’ expectations for its second quarter Meta’s shares rose in after-hours trading on Wednesday off the back of a strong earnings report that comes as the company is spending heavily on AI tools. The company’s stock price grew around 5% following the report, which revealed the company outperformed analysts’ expectations for its second quarter. Continue reading...