Posts with «author_name|mariella moon» label

Apple releases an AI model that can edit images based on text-based commands

Apple isn't one of the top players in the AI game today, but the company's new open source AI model for image editing shows what it's capable of contributing to the space. The model called MLLM-Guided Image Editing (MGIE), which uses multimodal large language models (MLLMs) to interpret text-based commands when manipulating images. In other words, the tool has the ability to edit photos based on the text the user types in. While it's not the first tool that can do so, "human instructions are sometimes too brief for current methods to capture and follow," the project's paper (PDF) reads. 

The company developed MGIE with researchers from the University of California, Santa Barbara. MLLMs have the power to transform simple or ambiguous text prompts into more detailed and clear instructions the photo editor itself can follow. For instance, if a user wants to edit a photo of a pepperoni pizza to "make it more healthy," MLLMs can interpret it as "add vegetable toppings" and edit the photo as such.

Apple

In addition to changing making major changes to images, MGIE can also crop, resize and rotate photos, as well as improve its brightness, contrast and color balance, all through text prompts. It can also edit specific areas of a photo and can, for instance, modify the hair, eyes and clothes of a person in it, or remove elements in the background. 

As VentureBeat notes, Apple released the model through GitHub, but those interested can also try out a demo that's currently hosted on Hugging Face Spaces. Apple has yet to say whether it plans to use what it learns from this project into a tool or a feature that it can incorporate into any of its products.  

This article originally appeared on Engadget at https://www.engadget.com/apple-releases-an-ai-model-that-can-edit-images-based-on-text-based-commands-081646262.html?src=rss

The EU wants to criminalize AI-generated porn images and deepfakes

Back in 2022, the European Commission released a proposal for a directive on how to combat domestic violence and violence against women in other forms. Now, the European Council and Parliament have agreed with the proposal to criminalize, among other things, different types of cyber-violence. The proposed rules will criminalize the non-consensual sharing of intimate images, including deepfakes made by AI tools, which could help deter revenge porn. Cyber-stalking, online harassment, misogynous hate speech and "cyber-flashing," or the sending of unsolicited nudes, will also be recognized as criminal offenses.

The commission says that having a directive for the whole European Union that specifically addresses those particular acts will help victims in Member States that haven't criminalized them yet. "This is an urgent issue to address, given the exponential spread and dramatic impact of violence online," it wrote in its announcement. In addition, the directive will require member states to develop measures that can help users more easily identify cyber-violence and to know how to prevent it from happening if possible or how to seek help. It will require them to provide their residents with an online portal where they can send in reports, as well. 

In its reporting, Politico suggested that the recent spread of pornographic deepfake images using Taylor Swift's face urged EU officials to move forward with the proposal. If you'll recall, X even had to temporarily block searches for the musician's name after the images went viral. "The latest disgusting way of humiliating women is by sharing intimate images generated by AI in a couple of minutes by anybody," European Commission Vice President Věra Jourová told the publication. "Such pictures can do huge harm, not only to popstars but to every woman who would have to prove at work or at home that it was a deepfake." At the moment, though, the aforementioned rules are just part of a bill that representatives of EU member states still need to approve. "The final law is also pending adoption in Council and European Parliament," the EU Council said. According to Politico, if all goes well and the bill becomes a law soon, EU states will have until 2027 to enforce the new rules.

This article originally appeared on Engadget at https://www.engadget.com/the-eu-wants-to-criminalize-ai-generated-porn-images-and-deepfakes-105037524.html?src=rss

Disney+ has started cracking down on password sharing in the US

Disney+ started getting strict about password sharing in Canada last year, and now it's expanding the restriction to the US. According to The Verge, the streaming service has been sending out emails to its subscribers in the country, notifying them about a change in its terms of service. Its service agreement now states that users may not share their passwords outside of their household "unless otherwise permitted by [their] service tier," suggesting the arrival of new subscription options in the future. 

The Verge says Disney+ told subscribers that they can analyze the use of their account to "determine compliance," though it didn't elaborate on how its methods work exactly. "We're adding limitations on sharing your account outside of your household, and explaining how we may assess your compliance with these limitations," Disney+ reportedly wrote in its email. In its Service Agreement, the service describes "household" as "the collection of devices associated with [subscribers'] primary personal residence that are used by the individuals who reside therein." The rule already applies to new subscribers, but old ones have until March 14 to feel its effects. 

Disney's other streaming service, Hulu, also recently announced that it's clamping down on password sharing outside the subscriber's "primary personal residence." It used the same language in its its warning to users, also telling them that their accounts will be analyzed for compliance and that it will start enforcing the new rule on March 14. 

This article originally appeared on Engadget at https://www.engadget.com/disney-has-started-cracking-down-on-password-sharing-in-the-us-070317512.html?src=rss

An Elden Ring mobile game is reportedly in the works at Tencent

Tencent is in the midst of developing an Elden Ring game for mobile phones, according to Reuters. Apparently, if the Shenzhen-based tech company's plan pushes through, fans are getting a free-to-play title with in-app purchases, similar to the gacha game Genshin Impact. That's a complete departure from the original Elden Ring action RPG, which sells for $60 from the get-go and isn't loaded up with microtransactions. 

That's pretty much all the details Reuters has reported — according to the news organization's sources, progress on the project has been slow, so it may take a while before we see the game. That is, if the game even gets released: Tencent recently shut down the development of much-awaited Nier mobile game that had been in the works for two years. Tencent reportedly couldn't find a monetization model that would allow it to recoup what it cost the company to license the franchise from Square Enix, along with what it's going to spend to bring the project to completion. 

Tencent and Sony purchased a 30 percent share of FromSoftware, the Japanese developer behind Elden Ring, back in 2022. The Chinese company also acquired the licensing rights to develop the game back then and put together a team with a "few dozen people" to work on a prototype. Reuters says Tencent is hoping to give its aging library of games a boost with a new title, and one based on a wildly successful game like Elden Ring has the potential to become a hit. 

This article originally appeared on Engadget at https://www.engadget.com/an-elden-ring-mobile-game-is-reportedly-in-the-works-at-tencent-131229933.html?src=rss

Sony's latest PS5 beta update gives DualSense controllers better audio

The latest beta update for the PlayStation 5 will make it easier to hear what's going on in your game. Sony Interactive Entertainment (SIE) has announced the software update, which will enable DualSense and DualSense Edge wireless controllers to produce louder audio so that you can hear in-game sounds and audio chats more clearly. 

In addition to giving controllers the ability to produce louder sounds, the update also gives their noise cancellation feature a boost. Sony's announcement says the controllers now have better mic input, thanks to a new AI machine learning model. It'll give PS5's system the ability to suppress background noises, such as the game audio itself, allowing your voice to go through clearly and improving the quality of voice chats as a whole. 

The update brings non-sound-related improvements to the PS5, as well, including the option to adjust the console's power indicator under "Beep and Light" in System Settings. Plus, it will give you access to Share Screen tools that you can use as a viewer to interact with the host's gameplay. Want to catch the host's attention and make them notice a certain element in the game? There's a pointer you can move around or use to draw a line on the display and highlight objects in game. You'll also be able to send the host's screen emoji reactions. These new interactive tools for the Share Screen are enabled by default, but you can always switch them off in settings. Take note that they will only work if both you and the host already have access to the beta features. At the moment, the upgrades are only rolling out to select users in certain countries, but SIE plans to release them to all PS5 gamers around the world in the coming months. 

This article originally appeared on Engadget at https://www.engadget.com/sonys-latest-ps5-beta-update-gives-dualsense-controllers-better-audio-113529372.html?src=rss

Apple Vision Pro users will have to go to a store to reset forgotten passcodes

Can you imagine having to mail in a $3,499 gadget because you forgot the passcode you nominated to be able to use it? Customers who purchased the Apple Vision Pro may have to do just that if they forget their device's passcode. According to Bloomberg, the company has been telling buyers that the only way to reset their codes is to go to an Apple Store or to mail their mixed reality headset to AppleCare customer support. Apple will then reset their codes so they could use their device again. 

As Bloomberg notes, if a user enters an incorrect code too many times, their device will be disabled and they'll have to sit through a waiting period to be able to try again. They'll have to get Apple to reset their code if they still can't recall it after that waiting period, and getting the company to do will erase all the content on their headset. 

It's worth noting that some users on the Apple discussion forum still got locked out even though they entered the correct passcode — it's unclear if it's a bug affecting some units — so the issue could happen to anyone. Based on some other posts, Apple's Genius Bar personnel use an accessory called the Developer Strap that the company is selling to developers for $299 to connect the headset to iPads or laptops to be able to reset it. In at least one instance, Apple Geniuses reportedly failed to reset the user's headset and had to replace it altogether. 

The process to change the headset's passcode could still change in the future — the recently released Vision Pro, after all, is the first iteration of a brand new product category. For the sake of those who've already purchased one, we hope it's introduced through a software update instead of as an improvement in the next version. 

This article originally appeared on Engadget at https://www.engadget.com/apple-vision-pro-users-will-have-to-go-to-a-store-to-reset-forgotten-passcodes-070101287.html?src=rss

NYC ends trial run of Times Square subway's security robot

New Yorkers can say goodbye — for now — to the robot the New York Police Department (NYPD) used to patrol the Times Square subway station. The Knightscope K5, which is a 400-pound machine that looks like it could be R2-D2's massive sibling, was one of Mayor Eric Adams' high-tech experiments for New York City. It was deployed in a pilot last year ostensibly to help keep commuters safe, but it had no actual capabilities that could be useful in situations that had to be addressed on site. "The K5 Knightscope has completed its pilot in the NYC subway system," a spokesperson for the department told The New York Times

The K5 has cameras that can record video and a button that people can press to get in touch with a live agent. It patrolled the station from midnight until 6AM throughout its trial run that lasted for two months, though "patrolled" might be a generous way of putting it. The machine has no arms and can't go up or down the stairs. Commuters told The Times that it usually has a couple of cops with it and that it almost always seemed to be plugged in. 

New York leased the K5 for around $9 an hour during its trial phase that ended in December. "This is below minimum wage," Adams said when he announced the robot's deployment at the Times Square station. "No bathroom breaks, no meal breaks." At the time, privacy advocates had been concerned that the machine could eventually be equipped with facial recognition and other surveillance tools. They no longer have to worry about the K5 being used for surveillance in NYC, unless, of course, city officials decide to keep on leasing it for other purposes. A Knightscope spokesperson told The Times that the company is "not authorized to speak about certain clients," which presumably includes the NYC government. 

This article originally appeared on Engadget at https://www.engadget.com/nyc-ends-trial-run-of-times-square-subways-security-robot-105528275.html?src=rss

DoorDash increases NYC delivery fees following new minimum wage rules

DoorDash customers in NYC will notice a new fee tacked onto their bill when they purchase food for delivery through the app. The company has started charging users $2 more for deliveries in NYC as a response to the city's new minimum wage law, a spokesperson has confirmed to Business Insider. It warned users back in December that the new minimum pay rate, which it called "ill-conceived" and "extreme," will have "significant consequences for everyone" who uses its platform and will "force [it] to raise fees for orders." Other major cities implementing a minimum pay rate for app-based deliveries will also be affected. Seattle customers, for instance, were recently hit with new fees worth 10 cents to $3.40 order. 

Under the new regulations, services like Uber, DoorDash and Grubhub will have to pay workers at least $18 an hour. DoorDash has chosen to pay drivers $29.93 for every active hour only, which means they're unpaid for the time they spend waiting for orders to come in. When the company published its response to Seattle's new rules, it said it was going to reduce the suggested tip amount for each purchase "in order to better balance the impact of these new costs and provide the best experience for consumers." 

Customers can still tip any amount they want, but they may be less inclined to tip as much as before due to the added fees. That's one possible direct impact to drivers, since as DoorDash notes in its announcement, they get 100 percent of customers' tips. That hasn't always been the case. Back in 2019, news reports exposed the company's practice of pocketing tips and using that money to pay for drivers' guaranteed fees, which should've come from DoorDash itself. The food delivery service only introduced a new earnings and tipping policy that ensures drivers are getting their tips shortly after those reports came out. 

This article originally appeared on Engadget at https://www.engadget.com/doordash-increases-nyc-delivery-fees-following-new-minimum-wage-rules-105051707.html?src=rss

Tesla settles California hazardous waste lawsuit for $1.5 million

Tesla and the 25 California counties that sued the automaker for mishandling hazardous waste at its facilities around the state have already reached an agreement just a few days after the lawsuit was filed. The court has ordered the automaker to pay $1.5 million as part of the settlement, which also includes hiring a third party to conduct annual waste audits of its trash containers for five years. These auditors will be taking a close look at the company's trash containers to check for hazardous materials. 

The counties that sued Tesla, which include Los Angeles and San Francisco, accused the company of dumping improperly labeled materials at transfer centers and landfills that were "not permitted to accept hazardous waste." Based on the complaint filed in San Joaquin County, Tesla was illegally disposing the waste it generated manufacturing and servicing its vehicles. 

Undercover investigators from the environmental division at the San Francisco District Attorney's Office were the first to find evidence of Tesla's illegal activities back in 2018. They found trash containers at the company's service centers containing materials, such as aerosols, antifreeze, lubricating oils, brake cleaners, lead acid batteries, aerosols, antifreeze, waste solvents, electronic waste and waste paint when they weren't supposed to. Investigators from other California counties' District Attorney's offices conducted their own investigations and found similar unlawful disposals. The Alameda country authorities who looked into its Fremont factory activities, for instance, discovered illegal disposal of waste containing copper and primer-contaminated debris. 

Tesla reached a settlement with the Environmental Protection Agency over its handling of hazardous materials back in 2019 and had to agree to properly manage waste at its Fremont plant in addition to paying a $31,000 fine. The automaker had also taken steps to screen its trash containers for hazardous waste before taking them to the landfill after being notified of the issue. But as District Attorney Brooke Jenkins said, "today's settlement against [the company] serves to provide a cleaner environment for citizens throughout the state by preventing the contamination of [their] precious natural resources when hazardous waste is mismanaged and unlawfully disposed." By having a third party regularly check whether Tesla continues to comply with the agreement, authorities can ensure that the company isn't illegally dumping harmful materials across the state over the next few years. 

This article originally appeared on Engadget at https://www.engadget.com/tesla-settles-california-hazardous-waste-lawsuit-for-15-million-070513014.html?src=rss

The FCC wants to make robocalls that use AI-generated voices illegal

The rise of AI-generated voices mimicking celebrities and politicians could make it even harder for the Federal Communications Commission (FCC) to fight robocalls and prevent people from getting spammed and scammed. That's why FCC Chairwoman Jessica Rosenworcel wants the commission to officially recognize calls that use AI-generated voices as "artificial," which would make the use of voice cloning technologies in robocalls illegal. Under the FCC's Telephone Consumer Protection Act (TCPA), solicitations to residences that use an artificial voice or a recording are against the law. As TechCrunch notes, the FCC's proposal will make it easier to go after and charge bad actors. 

"AI-generated voice cloning and images are already sowing confusion by tricking consumers into thinking scams and frauds are legitimate," FCC Chairwoman Jessica Rosenworcel said in a statement. "No matter what celebrity or politician you favor, or what your relationship is with your kin when they call for help, it is possible we could all be a target of these faked calls." If the FCC recognizes AI-generated voice calls as illegal under existing law, the agency can give State Attorneys General offices across the country "new tools they can use to crack down on... scams and protect consumers."

The FCC's proposal comes shortly after some New Hampshire residents received a call impersonating President Joe Biden, telling them not to vote in their state's primary. A security firm performed a thorough analysis of the call and determined that it was created using AI tools by a startup called ElevenLabs. The company had reportedly banned the account responsible for the message mimicking the president, but the incident could end up being just one of the many attempts to disrupt the upcoming US elections using AI-generated content. 

This article originally appeared on Engadget at https://www.engadget.com/the-fcc-wants-to-make-robocalls-that-use-ai-generated-voices-illegal-105628839.html?src=rss