OpenAI's safety committee to oversee security practices as independent body

OpenAI's safety committee to oversee security practices as independent body


OpenAI, the company behind the viral chatbot ChatGPT, formed its Safety and Security Committee this May to evaluate and further develop the company’s existing safety practices | Photo: Bloomberg


Microsoft-backed OpenAI said on Monday its safety committee will oversee security and safety processes for the company’s artificial intelligence model development and deployment, as an independent body.

 


The change follows the committee’s own recommendations to OpenAI’s board which were made public for the first time.

 


OpenAI, the company behind the viral chatbot ChatGPT, formed its Safety and Security Committee this May to evaluate and further develop the company’s existing safety practices.

 


The launch of ChatGPT in late 2022 sparked a significant wave of interest and excitement around AI.

 

 


The buzz around ChatGPT highlighted both the opportunities and challenges of AI, prompting discussions on ethical use and potential biases.

 


As part of the committee’s recommendations, OpenAI said it is evaluating the development of an “Information Sharing and Analysis Center (ISAC) for the AI industry, to enable the sharing of threat intelligence and cybersecurity information among entities within the AI sector.”

 


The independent committee will be chaired by Zico Kolter, professor and director of the machine learning department at Carnegie Mellon University, who is part of OpenAI’s board.

 


“We are pursuing expanded internal information segmentation, additional staffing to deepen around-the-clock security operations teams,” according to OpenAI.

 


The company also said it will work toward becoming more transparent about the capabilities and risks of its AI models.

 


Last month, OpenAI signed a deal with the United States government for research, testing and evaluation of the company’s AI models.


(Only the headline and picture of this report may have been reworked by the Business Standard staff; the rest of the content is auto-generated from a syndicated feed.)

First Published: Sep 17 2024 | 10:51 AM IST



Source link

HMD Skyline with Nokia Lumia-inspired design launched in India: Know more

HMD Skyline with Nokia Lumia-inspired design launched in India: Know more



Finnish mobile phone brand Human Mobile Devices (HMD) has launched the Skyline smartphone in India. Unveiled in July, the smartphone features a design inspired by the Nokia Lumia series and emphasises repairability. HMD states that the Skyline smartphone simplifies at-home repairs through the quick and easy disassembly of key replaceable components.


HMD Skyline: Price and availability


The HMD Skyline is offered in a 12GB RAM and 256GB on-board storage configuration. Priced at Rs 35,999, the smartphone is now available in India on HMD official website, the e-commerce platform Amazon India, and at select retail outlets. The HMD Skyline comes in two colours: Twisted Black and Neon Pink.

 


HMD Skyline: Details


Powered by the Qualcomm Snapdragon 7s Gen 2 chip, the HMD Skyline features a 6.55-inch fullHD+ pOLED display of a 144Hz refresh rate. The display supports HDR10 for compatible content. For imaging, the smartphone has a 108MP primary camera with optical image stabilisation. This primary sensor is accompanied by a 50MP telephoto camera with 4x optical zoom and a 12MP ultra-wide angle camera.


The smartphone is powered by a 4600 mAh battery and supports 33W wired charging. Additionally, it supports wireless charging with magnetic attachment, enabled by Qi2 support.


Regarding repairability, HMD states that the Skyline smartphone employs an internal screw-driven camshaft that allows users to easily disassemble the device at home. By turning a single screw, users can remove the back cover. They can also disconnect the display cable and replace the removable battery independently.


HMD Skyline: Specifications


  • Display: 6.55-inch pOLED, FHD+, 144 Hz refresh rate, HDR10

  • Processor: Qualcomm Snapdragon 7s Gen 2

  • RAM: 12GB

  • Storage: 256GB

  • Rear Camera: 108MP primary (OIS) + 50 MP telephoto (4x zoom) + 12 MP ultra-wide

  • Front Camera: 50MP (auto focus)

  • Battery: 4600 mAh

  • Charging: 33W wired, Qi2 wireless charging

  • Operating System: Android 14 (two OS updates + three years of security updates)

First Published: Sep 17 2024 | 10:40 AM IST



Source link

Social media age limits restricting children are too little too late

Social media age limits restricting children are too little too late


Australia is taking matters into its own hands | Photo: Bloomberg


By Catherine Thorbecke

 


Australia’s government wants to ban children up to age 16 from social media, and is spending millions of dollars to figure out how. I’m willing to wager it won’t take long for tech-savvy teens who grew up on Instagram, TikTok and YouTube to figure out how to log back on.

 


The promised regulation, currently sparse on details, comes at a time when policymakers and parents around the globe are grappling with the negative consequences these platforms can have on developing minds. This global debate has raged for years, reaching a fever pitch in 2021 after former Facebook (now Meta Platforms Inc.) employee Frances Haugen leaked documents showing the company was aware its products were harmful to girls’ mental health. Years later, US lawmakers are still sputtering on federal regulation to keep the powerful Big Tech companies accountable for harms to young users. 

 


Australia is taking matters into its own hands. Prime Minister Anthony Albanese promised to introduce new laws that set age limits this year, saying that the government was considering a range between 14 and 16 for the cutoff. In a video posted on X for “the mums and dads,” Albanese said he wants children “off their devices and onto the footy field.” Surveys indicate most Australians support a social-media age limit, and the idea has broad political support.


But even Albanese acknowledges that they are still trying to figure out how this would actually work. The government doesn’t identify what social media platforms the youth ban would apply to (Can children message their parents on WhatsApp? Or watch Khan Academy’s Algebra tutorials on YouTube?). It also doesn’t offer specifics on enforcement (Big Brother-esque digital IDs? Further criminalizing children, this time for opening TikTok?). And in the absence of substantive policies, it’s hard not to see this as a soundbite-y proposal to signal concern to voting  parents on a popular issue ahead of an election year — without actually accomplishing anything to keep children safe. 


Thousands of miles away from Silicon Valley, Australia has been leading the charge in efforts to rein in the dominance of Big Tech. Separate proposed legislation aimed at cracking down on digital misinformation has even drawn ire from Elon Musk, who last week labeled the government “fascists.” (The government has sued Musk’s X, formerly known as Twitter, over a violent video of a terrorist attack but lost in court.) The nation has also been engaged in a years-long battle to force tech titans to pay for news content. At a time when other jurisdictions have struggled with taking on such powerful companies, Australia’s multi-faceted attacks are admirable.


But research has shown that age limits for social media aren’t the most effective way to protect teens from its potential harms. Young people have shown remarkable prowess for finding workarounds — even those under the age of 13 whom most platforms already prohibit. The American Psychological Association has argued that using social media is not inherently beneficial or harmful to teens, but strict age limits ignore individual differences in adolescents’ maturity levels. In other words, turning 16 doesn’t instantly make you more competent at navigating the digital world than a mature 14-year-old.


The process of enforcing broad age verification online raises a slew of privacy concerns, ranging from how identifying information about young users could be stored to cutting off their ability to freely browse the internet while maintaining digital anonymity.


Completely shutting off access to digital communities can also sever lifelines for some young people, especially those from marginalized groups. TikTok, in particular, has emerged as a popular platform for Indigenous Australians, allowing them a space where they share everything from budget-friendly recipes to relatable responses to racism. Indigenous youth in remote areas who may not see their stories reflected in traditional media can feel less isolated. LGBTQ+ advocates in Australia have raised similar concerns about a potential loss of connections for vulnerable queer teens if the ban takes effect. More broadly, tech researchers warn that excluding young people from social media platforms will just drive them to darker, even less regulated corners of the web. 


Still, a growing body of evidence points to a minefield of harms young people can encounter, as much as company executives like to deflect any links. It’s absolutely critical that lawmakers take action to protect children from these risks, but selling quick fixes for complex, global problems distracts from the harder policy work required to come up with effective real-world solutions. 


Simply banning young people from participating in digital life comes a generation too late. The reality is teens today are very much growing up online, a trend accelerated by the pandemic. So much so that the United Nations has said that children have the right to get information from the internet, but adults have a responsibility to make sure it isn’t harmful. 


Policymakers need to focus on holding social media companies accountable for the harms, especially for young users, embedded within their services. They can start by demanding that platforms offer more transparency about how their algorithms work and allowing more outside researchers to look under the hood to identify risks. Without sharing data on how their services are designed, it’s hard for mental health experts and officials to recommend solutions that address the dangers. Lawmakers must also focus on requiring social media companies, which go to great lengths to understand their users, to create and enforce more guardrails for young people.  


Without putting the onus on tech companies to reduce risks on their platforms, raising the age limit by a couple of years doesn’t keep the next generation safe. Instead of bucketing out floodwater, policymakers in Australia and beyond should turn off the spewing faucets.


Disclaimer: This is a Bloomberg Opinion piece, and these are the personal opinions of the writer. They do not reflect the views of www.business-standard.com or the Business Standard newspaper

First Published: Sep 17 2024 | 9:47 AM IST



Source link

US justice dept says Google saw ad startup AdMeld as threat and bought it

US justice dept says Google saw ad startup AdMeld as threat and bought it


Google purchased AdMeld for more than $400 million and, after integrating its technology into the company’s advertising exchange, shut down the product two years later | Photo: Bloomberg


By Leah Nylen and Davey Alba

 


Google bought advertising technology provider AdMeld in 2011 because the search giant saw the startup as a “threat” to its online display ad strategy, antitrust enforcers sought to show at trial Monday.

 


The purchase of the company was portrayed by the Justice Department in its monopolisation trial against the Alphabet Inc. unit as an example of so-called killer acquisitions. The government alleges that Google bought up nascent rivals, starting with DoubleClick in 2008 and followed by AdMeld and Invite Media, to build up a dominant position and stifle competition for its online ad tools.

 


After purchasing DoubleClick, maker of the web’s leading ad server, Google employees debated whether to buy a company that made “yield management tools,” which help websites analyze data for ad pricing. Leading yield management companies at the time included AdMeld, PubMatic Inc. and the Rubicon Project. 


In an internal presentation shown in court Monday, Google staff dismissed the technology as “irrelevant,” but said the tools were getting in the way of the search giant’s opportunities to include more website ad inventory on its ad exchange platform. 


Google purchased AdMeld for more than $400 million and, after integrating its technology into the company’s advertising exchange, shut down the product two years later.


The Justice Department and a group of states allege that Google has monopolised the market for advertising technology tools used by websites and advertisers to buy and sell online display ads. Antitrust enforcers are seeking to show that Google bought rival companies to solidify its power in the online ad market. 


In an October 2010 email to a colleague, YouTube CEO Neal Mohan, then a leader at Google’s display advertising division, wrote that the company had “missed the yield manager threat.”


Mohan, who joined the search giant when it bought DoubleClick in 2008, suggested in a message to colleagues that Google should buy one of the leading products by “picking up the one with the most traction and parking it somewhere.”


In court on Monday, Mohan denied that he was suggesting Google buy AdMeld to eliminate a competitor.


“Absolutely not,” he said. AdMeld’s technology “was a gap in our portfolio.”


‘Close that gap’

 


“We needed to close that gap as quickly as possible,” Mohan added.


In the same presentation about whether the company should acquire a yield manager, Google said that buying AdMeld or PubMatic would close “product and service gaps” and was “a pressing need for publishers.” An acquisition would also “ensure fair access to publisher inventory.”


In his testimony, Mohan said that AdMeld was a “complement” to Google’s advertising exchange, AdX, a platform which matched advertisers with websites that wanted to sell space.


“Yield management did something adjacent to, but different than, what AdX did,” he said.


Mohan said that in his view Google’s products were more advanced than yield management tools, likening AdX to streaming video while AdMeld and PubMatic were DVDs aimed at “solving yesterday’s problems.” But many websites were “more cautious” about taking the leap to Google’s products, Mohan said, and were more comfortable adopting tools like AdMeld and Pubmatic, which he described as “baby steps” away from the technology they were used to.


At the same time, Mohan acknowledged that Google’s acquisition of both DoubleClick and eventually AdMeld was done in order to keep the technology giant from “falling behind” on its ad-tech offerings.


“If we lose platform share, we can build the best” advertising platform “in the world but will still be at a severe risk of being disintermediated” by large competitors, Mohan wrote in a March 2009 email to colleagues, referring to companies such as Yahoo Inc. and Microsoft Corp., which were working on building up their own ad offerings. In the same email, he later added that Google needed “tight bundles” in its ad offerings that it could offer both publishers and advertisers in order to stay competitive.


Google employees estimated that AdMeld was worth between $182 million and $355 million. Mohan acknowledged that Google ultimately paid $100 million above what the search giant estimated AdMeld’s value was.

First Published: Sep 17 2024 | 9:15 AM IST



Source link

Ms signs office AI deal with Vodafone, adds features to Excel, Outlook

Ms signs office AI deal with Vodafone, adds features to Excel, Outlook


By Dina Bass and Matt Day

Microsoft Corp. said Vodafone Group Plc signed a large deal to use artificial intelligence assistants for Office and unveiled a range of new AI tools designed to help workers create Excel charts, prioritize emails and collaborate with one another.


Vodafone will purchase 68,000 licenses for corporate versions of the assistants, known as Copilots, after early trials found workers using them saved about three hours a week per person, Microsoft said in a blog post shared ahead of a webcast Monday. Other clients include Honeywell International Inc. and financial services firm Finastra Group Holdings Ltd. Microsoft said Copilot customers increased by more than 60 per cent in the most recent quarter, compared with the previous period, while the number of clients with more than 10,000 users doubled. The company didn’t disclose overall customer totals. 

 


Microsoft has been infusing its product line with AI technology from partner OpenAI and is touting the new features as a second wave of business-oriented Copilots, AI tools designed to turbocharge programs like Teams conferencing software, Word and Outlook, among other applications. 


Connecting popular Office software to AI models from OpenAI helps differentiate Microsoft’s products from rival offerings, including OpenAI’s own. OpenAI last week released a new model, called o1, that can perform some human-like reasoning tasks, and Microsoft said Monday that it would integrate the technology in its own products at some point.


Microsoft has cited strong demand for the earlier versions of the products, and customers praise some of their capabilities, especially the ability to summarize meetings and documents. The Copilots cost $30 a month per user and could one day become a major new revenue stream.

But many features remain a work in progress — generating new content has been more unreliable and the same is true of using the AI tools for things like spreadsheets or tasks that require understanding context. In recent months, skittish investors have signaled impatience with tech companies’ efforts to profit from their massive investments in AI.


Microsoft is also rolling out a new feature for Copilot’s chat application called Pages, which gives coworkers a space to work together using data they create and as well as content generated by the AI software. Pages saves the AI-generated material in one place and lets workers edit and share it with other employees. The new feature is being released Monday and will be widely available later this month. “You can ideate with AI and collaborate with other people,” said Microsoft Chief Executive Officer Satya Nadella. 


Microsoft is also widely releasing its Copilot tools for the Excel spreadsheet app, which will help users create charts and pivot tables. The Excel Copilot can also work with text data and use the Python programming language for things like forecasts and risk analysis. The Python features, in preview now, are accessible using regular language instead of programming code. 


The software giant announced new AI features for Word, Teams and PowerPoint. A new Prioritize My Inbox option in Outlook analyzes a user’s messages based on what emails they respond to and who they report to in the office in order to help declutter. The program, which will be available to preview late this year, also summarizes the messages and explains why they’ve been prioritized. 

First Published: Sep 17 2024 | 12:03 AM IST



Source link

Tech wrap Sep 16, Tech News: iOS 18 rollout, Lava Blaze 3, Motorola Edge 50 Neo, more

Tech wrap Sep 16, Tech News: iOS 18 rollout, Lava Blaze 3, Motorola Edge 50 Neo, more



Apple will start rolling out the iOS 18 operating system to eligible iPhone models today, September 16. Although the new iPhone 16 series, which comes with iOS 18 pre-installed, will be available from September 20, older models that qualify for the update will begin receiving it starting today.


Indian smartphone brand Lava has unveiled the Blaze 3 5G. This device, equipped with the MediaTek Dimensity 6300 5G chip, boasts a glass back, dual stereo speakers, and a new “Vibe Light” feature designed to enhance photography with a studio-like effect. The Blaze 3 5G is available in two colours: Glass Gold and Glass Blue.

 


Motorola, a smartphone brand owned by Lenovo, is growing its Edge 50 series in India with the introduction of the Motorola Edge 50 Neo. According to Motorola, this new model offers US Military Grade certification for durability and an IP68 rating for protection against dust and water. The smartphone includes a 50-megapixel main sensor (Sony LYTIA) and a 3x telephoto camera.


Xiaomi has launched the Redmi Smart Fire TV 4K series in India, featuring two models: a 55-inch and a 43-inch version. Both models come with 2GB of RAM and 8GB of storage. They provide a 4K resolution display and support High Dynamic Range (HDR) for improved picture quality.


According to analyst Ming-Chi Kuo, Apple’s new iPhone 16 series seems to be experiencing lower-than-expected demand. In a report on Medium, Kuo observed that pre-order sales for the iPhone 16 series during the first weekend were 12.7% lower compared to the iPhone 15 series. The drop is especially notable for the iPhone 16 Pro models.


Apple has confirmed that the new iPhone 16 and iPhone 16 Plus models come with 8GB of RAM, an increase from the 6GB RAM found in last year’s base models. Johny Srouji, Apple’s senior vice president of hardware technologies, revealed this information during an interview with Chinese YouTube creator Geekerwan.


Apple is shipping the iPhone 16 series retail units without stickers. As detailed in a recent memo to Apple Store teams, reported by 9to5Mac, the iPhone 16 packaging will no longer include the sticker. In-store buyers can request an Apple sticker if they wish, but those who order online for home delivery will not receive one. Additionally, third-party retailers and carrier partners will not have Apple stickers available.


Apple is releasing the new iOS 18 operating system today, September 16. However, the initial rollout will not include Apple Intelligence features. These advanced AI tools will be introduced in later updates over the next few months. The upcoming iOS 18.1 update, expected by October, will feature initial Apple Intelligence tools like text summarization, text generation, and notification summaries. More advanced functionalities, including Image Playground, Genmoji, and ChatGPT integration, are anticipated with iOS 18.2 in December.


Google’s gesture-based Circle to Search feature, which was previously limited to Pixel and Samsung devices, is expected to expand to more Android smartphones. According to a report by Android Authority, TECNO, a Chinese smartphone manufacturer, has notified YouTuber Ben Sin and Android Central that the feature will be available on its upcoming foldable device, the Phantom V Fold 2, beginning next month.


Meta is reportedly preparing to introduce a new feature enabling users to share comments from Instagram posts to Threads. According to The Verge, app researcher Alessandro Paluzzi has shared an image indicating that a dropdown menu will appear when users comment on an Instagram post. This menu will provide the option to share the comment solely on Instagram or also on Threads.


Apple is reportedly contemplating the use of a rigid plastic casing for the next-generation Apple Watch SE, potentially set for release next year. This move comes after Apple’s previous experimentation with plastic, which was not featured in recent product announcements.


Intel missed out on a contract to design and manufacture the chip for Sony’s PlayStation 6 in 2022, according to three sources familiar with the situation. This loss was a substantial setback for Intel’s efforts to grow its emerging contract manufacturing business.

First Published: Sep 16 2024 | 8:03 PM IST



Source link

YouTube
Instagram
WhatsApp