Roblox is preparing to roll out its biggest change since starting age verification late last year. While that program was initially focused on chat access, today’s news is about age-segregating the games on the massive platform. Starting in mid-May, users will be pushed into one of three worlds: Roblox Kids, Roblox Select or Roblox. The exact age ranges of these groups will vary by territory, but in the US they are 5-8 for Kids, 9-15 for Select and 16+ for the regular account. These three account types then align with the platform’s current content maturity labels, which divide games into Minimal, Mild, Moderate and Restricted. Kids accounts will be the most restricted, with chat off by default and only Minimal and Mild experiences available. Ages 9-15 get to chat with kids in their age group and “trusted friends” that have passed the parent test, and will be able to access Moderate content as well as games for babies. At 16, teens will automatically be moved to a full-fat Roblox account with all of its features, but not all of the games. Content marked as Restricted will only unlock once they turn 18. Roblox Roblox says over half of its users are now age checked, whether through ID verification or face scans. With the new account types rolled out globally — which the company says should be done by June — it’ll start forcing users who haven’t completed an age check into a Kids-like experience, with no access to chat or games rated higher than Mild. Once age verification is completed, Roblox still faces the task of ensuring that its vast collection of user-created content is actually age-appropriate. Its solution to this is, of course, ID verification, AI and upcharges. Developers will have to verify their identity and pony up $5 a month for Roblox Plus to show “a long-term commitment to the platform.” The wisdom is that, with these hurdles cleared, a developer will surely apply the correct maturity label to their games. On the off-chance that an experience is mislabeled, Roblox’s AI moderation will keep tabs on game instances to make sure what's happening on-screen and in-chat matches the maturity label. On the surface, this does leave a gap where a toddler could end up playing an incorrectly labeled mature game before the AI catches it. Don’t fret, though, as Roblox says users over 16 “play new games first,” which surely isn’t an overgeneralization and will ensure that no child ever plays a mature game. Roblox also previewed a pair of new parental control features coming in June. First, parents will be able to block any game and manage direct chat access until a child turns 16. Previously, kids over 13 could unblock experiences by themselves. Second, parents will be able to approve games outside of their child’s age bracket on a case-by-case basis. Roblox gave an example of a younger child wanting to play a game with their older sibling for this feature’s utility. Of course, the big blocky elephant in the room is the efficacy of automated age verification. Reporting from Wired in January suggested even enterprising toddlers might be able to get past the platform’s age checks, which somewhat undermines everything Roblox is trying to achieve. Speaking to press ahead of today’s announcement, Roblox Chief Safety Officer Matt Kaufman said, “If we get it wrong … we offer users multiple ways to correct that.” He added that the platform is “constantly measuring users’ behavior and comparing that against what their age-check data says. If we see those things divert, then we will just ask people to run through the age process again.” This article originally appeared on Engadget at https://www.engadget.com/gaming/roblox-introduces-mandatory-age-gated-account-tiers-121535702.html?src=rss
OpenAI CEO Sam Altman's house may have been the target of a second attack after San Francisco Police Department arrested two suspects for a reported shooting in the Russian Hill neighborhood. The SFPD said in a press release that police officers responded to a "suspicious occurrence of possible shots fired" at around 5:56 AM ET / 2:56 AM PT on Sunday, April 12. SFPD's Special Investigation Division took over the case and have since detained both 25-year-old Amanda Tom and 23-year-old Muhamad Tarik Hussein, seizing three firearms in the process with the help of a warrant. The two suspects were charged with negligent discharge. According to the initial police report, as reported by The San Francisco Standard , two people inside a Honda sedan stopped in front of Altman's property that spans from Chestnut Street to Lombard Street. The police report also noted that the passenger appeared to fire a round at the Lombard Street side of Altman's property. The property's security personnel reported hearing a gunshot and there was surveillance footage that recorded the incident, according to the report. This could be the second instance of violence targeting Altman and his residence in a matter of days. On Friday, a 20-year-old man allegedly hurled a Molotov cocktail at Altman's home, which caused a fire on one of the property's exterior gates, according to SFPD. The San Francisco Standard reported that there were no injuries in either incident. This article originally appeared on Engadget at https://www.engadget.com/ai/two-suspects-have-been-arrested-for-allegedly-shooting-at-sam-altmans-house-142655308.html?src=rss
Picture this: You're a senior Meta employee looking for feedback from the CEO. But, instead of hearing from the real Mark Zuckerberg, you get a response from a Zuckerberg AI character. As absurd as that sounds, it could eventually be a reality. Meta is reportedly working on such an AI character, training it on Zuckerberg's mannerisms, tone and publicly available statements, according to the Financial Times . The character is also learning about the CEO's thoughts on recent company strategy, with the idea that it could offer advice to Meta employees. The company has reportedly, for some time, been working on creating photorealistic, 3D animated AI characters that can manage interactions. However, it now appears to be focusing on this Zuckerberg AI character, which would interact with employees when the CEO can't or doesn't want to. This additional AI tool follows last month's news that Zuckerberg is creating an AI agent to help him do his job, first reported by the Wall Street Journal . It would reportedly do things like finding answers for him, but there aren't many details of the still developing AI agent. This article originally appeared on Engadget at https://www.engadget.com/ai/meta-is-reportedly-building-an-ai-clone-of-mark-zuckerberg-130242840.html?src=rss
