Technology

GulfNews Technology

Facebook doubles down on detecting deepfakes

Media|: San Francisco: Facebook has collaborated with researchers at the Michigan State University (MSU) to develop a method of detecting and attributing deepfakes. It relies on reverse engineering, working back from a single AI-generated image to the generative model used to produce it. "Our reverse engineering method takes image attribution a step further by helping to deduce information about a particular generative model just based on the deepfakes it produces," said research scientists Xi Yin and Tal Hassner at Facebook. It's the first time that researchers have been able to identify properties of a model used to create a deepfake without any prior knowledge of the model. Deepfakes are being treated as video forgeries that make people appear to be saying things they never did, like the popular forged videos of Facebook CEO Zuckerberg and that of US House Speaker Nancy Pelosi that went viral. Deepfakes have become so believable in recent years that it can be difficult to tell them apart from real images. Image attribution can identify a deepfake's generative model if it was one of a limited number of generative models seen during training. "During image attribution, those deepfakes are flagged as having been produced by unknown models, and nothing more is known about where they came from, or how they were produced," said Facebook. The company said that with the new method, researchers will now be able to obtain more information about the model used to produce particular deepfakes. "Our method will be especially useful in real-world settings where the only information deepfake detectors have at their disposal is often the deepfake itself," Facebook said. To combat the spread of disinformation, Microsoft also last year unveiled a new tool that will spot deepfakes or synthetic media which are photos, videos or audio files manipulated by Artificial Intelligence (AI) which are very hard to identify if false or not.

GulfNews Technology

Google is totally changing how ads track people around the Internet. Here's what you need to know.

Technology|: It's a common sight: Ads from that time you Googled flights to Cancun, or visited Nike to look for new running shoes, following you around the Internet. Much of that tracking is made possible by cookies - little bits of code that jump off websites and lodge themselves in your browser, allowing new sites you visit to see where you've been before. Facebook and Google, the two most profitable advertising companies in history, use cookies to show ads across the Web based on info gathered on their own sites and social media networks. But that's all changing. Google has vowed to block cookies completely on its Chrome browser, which is used by around 70% of the world's desktop computer owners, by the beginning of 2022. The decision, announced last year, sent shock waves through the advertising world, which has maintained revenue from tracking is necessary to fund a largely free Web. Google says it has solutions to allow advertisers to keep showing relevant ads, but in privacy-protecting ways. Taken together, the company's proposals are meant to let Web publishers, e-commerce companies and advertising agencies continue using targeted ads to make money, while assuring regular Internet users their data isn't being stockpiled by an ever-growing list of companies and websites. But privacy activists have already started poking holes in Google's ideas. And it may not matter. Advertising technology companies such as the Trade Desk have already taken the matter into their own hands, banding together to create new tracking tools that use email addresses. Other major companies have shown signs of pushing back against Google's proposals, such as Amazon, which is currently blocking Chrome from collecting data on which users go to its websites. (Amazon chief executive Jeff Bezos also owns The Washington Post.) Meanwhile, politicians and antitrust investigators in multiple countries have raised alarms that Google's move could hurt competitors and further cement its power. And for regular Internet users, this largely behind-the-scenes change could have major implications for how private companies hoover up our data and make decisions about what we see online. Here's what you need to know. How did we get here? Cookies were written into early browsers to cut down on some of the inconveniences of surfing the Web. They allowed passwords to auto-fill, or websites to remember payment information so users didn't have to type theirs in every time they came back. They also created a trail of breadcrumbs that the burgeoning online ad industry eagerly ate up, helping free websites make money. But as the technology advanced, social media took off and consumers' lives were lived increasingly online, it got creepy. Privacy advocates have always criticized the model, and more and more regular people have become aware of the issue, some expressing their displeasure by downloading ad blockers. Google isn't the first to make this change. Apple in 2017 started limiting and eventually blocking third-party cookies completely from its Safari browser. Mozilla's Firefox followed soon after. But those two browsers make up less than 20 percent of the market, according to research firm eMarketer. Despite Google's own reliance on advertising and tracking for roughly $180 billion a year in revenue, chief executive Sundar Pichai admitted during a 2019 congressional hearing that people don't like to feel they're being tracked around the Internet. And in January 2020, Google said it too would block third-party cookies on Chrome within the next two years. The changes come as politicians in the United States and elsewhere step up their attempts to regulate privacy. The European Union's General Data Protection Regulation has forced companies to ask permission before tracking people online since 2018. In 2020, America's most populous state instituted the California Consumer Privacy Act, which gives California residents the right to ask companies to delete whatever data has been gathered on them. As is the case with other consumer-focused regulation, the California law has essentially become the default nationwide. Google was facing pressure from its competitors, too. Apple has been marketing its own privacy features aggressively, trying to paint itself as a privacy champion that doesn't need to gather data to feed an advertising business like Google. It even threw up a giant billboard that loomed over Google's exhibit at the 2019 CES tech conference in Las Vegas. Apple does gather some of its users data and uses it to sell targeted ads in its app store, though its ad business is much smaller than Google's. Some of Google's advertising technology competitors say the move isn't about privacy at all, but a way to hurt its rivals and push advertisers toward Google's YouTube and search ads, which don't need cookies to effectively target people. "You can fix your public perception while at the same time cementing your own dominance and growing your own market share," said Ratko Vidakovic, founder of AdProfs, an independent advertising technology consulting firm. "It seems like a no-brainer." A Google spokeswoman pointed to a company blog from March, where Marshall Vale, a product manager, said the company's goal with FLOC and its other projects is to make cookies obsolete while also helping web publishers grow their businesses. Finding that balance is "critical to keep the web open, accessible and thriving for everyone," Vale said. How exactly does Google's solution for the post-cookie world work? Google can block cookies in Chrome relatively easily because it designs and controls the browser's underlying code. Once it decides to make the change, it can update the browser and poof - no more cookies. To replace that functionality, Google's engineers have marched out a menagerie of bird-themed acronyms like FLOC, FLEDGE and TURTLEDOVE to describe their proposals for advertising without cookies. The ideas are working their way through the World Wide Web Consortium, or W3C, an international group of tech companies that debates and sets rules for how the Web works. However, Google doesn't actually need to get approval from the rest of the W3C's membership. Since its browser is the biggest in the world, it can simply make new rules and Web developers will have to follow them or risk seeing their websites stop working on Chrome. "Google using the W3C letterhead to do this stuff makes it seem less like a Google power play," said Peter Snyder, senior privacy researcher at Brave, a browser that competes with Google's Chrome. The most fleshed-out idea so far is FLOC, which stands for Federated Learning of Cohorts. Under FLOC, instead of letting websites drop cookies into an individual's browser, the browser itself watches what they do online. It then uses artificial intelligence to assign them to a cohort of several thousand people that the AI determines are interested in the same kinds of products. Then, instead of buying access to individual people, advertisers pay for ads to show up for users in a specific cohort. For example, if you spent the past few days reading articles on ESPN, browsing New York Knicks jerseys and Googling NBA stats, you might be lumped into a package of several thousand basketball fans who would see similar ads. Cohort IDs refresh every week, so they're based on the most recent browsing behavior. In the old world, websites would constantly be pulling up information about you based on the cookies trailing behind you. Now, the only identifying information your browser would present is which cohort you're in. Google says this system is 95 percent as effective at getting clicks as old-school cookie ads are for advertisers. If that's true, consumers would see pretty much the same kind of ads they do now and will probably still have the feeling of being followed around the Web by ads for sites they recently visited. Generally, yes, but that doesn't mean privacy advocates are celebrating the change. For one, Google's Chrome browser is still monitoring every website you visit and feeding that into its algorithm. The information stays on your device, but it's still being gathered. For those who want less surveillance from tech companies, it might feel like a step in the wrong direction. "The technology will avoid the privacy risks of third-party cookies, but it will create new ones in the process," Bennett Cyphers, a researcher with the Electronic Frontier Foundation, wrote in a March report on Google's cookie replacements. "It hasn't learned the right lessons from the ongoing backlash to the surveillance business model." It also isn't clear yet which websites will have access to a person's cohort ID. If it's freely available, sites you visit repeatedly could collect them as they change week to week, tie it to other pieces of information about you such as your email or IP address, and build a dossier on your interests, circumventing the stated purpose of FLOC, Cyphers argues. Google acknowledges this issue and says it is one of the long-term problems it is working on. The system also raises the possibility of profiling based on race, allowing advertisers to discriminate against some people. Advertising jobs or housing selectively by race is illegal in the United States. Still, compared to other proposals from the rest of the ad-tech industry, Google's is arguably the best one for privacy, Vidakovic said. "They're trying to balance commercial needs with user privacy needs at the same time," he said. "Despite their flaws, I think the concept behind FLOC and anonymous cohorts are a good balance." What does Google's move mean for competition? Unlike FLOC, cookies aren't owned and controlled by a specific company. They are a generic technology that any Web publisher or ad-tech seller can use to track people and show them ads. The world of cookie advertising resembles a capitalist Wild West, where anyone can hang a shingle and try building a fortune in Web ads. Google's new FLOC system is more controlled, laying out strict rules for how exactly advertisers can interact with the people who use Chrome. Cookies have also been used extensively to check how effective digital ads are. With FLOC, advertisers would have to trust Google that the ads they're paying for are being shown to the right people. Competitors to Google argue the company is pulling the ladder up behind it. Google used cookies to help it build a massive advertising business, but because YouTube and Google Search - which don't need cookies - are its biggest moneymakers, it can afford to live in a cookie-free Web. Advertisers who can't use cookies to find people on the open ocean of the Web will give more of their money to Google and Facebook who can pinpoint the right targets on their own sites, which industry insiders call "walled gardens." In January, the UK's competition authority said it would investigate FLOC and Google's other ideas to "assess whether the proposals could cause advertising spend to become even more concentrated on Google's ecosystem at the expense of its competitors." On the flip side, if Google were to simply shut down third-party cookies without building an alternative like FLOC, small companies, and the consumers who look to them for innovative new products, could pay the price. Big brands who already have contact information for their customers can use email marketing to reach them, while start-up retailers use targeted ads to find new people. Without targeted ads, companies such as glasses seller Warby Parker or makeup start-up Glossier might never have survived long enough to compete and bring down the prices that older companies were charging consumers. The same dynamic applies to publishing. Big news organizations who have paying subscribers don't rely as much on targeted ads as small, local news providers. If those small news providers have even fewer ways to make money, the communities they serve will suffer. (The Washington Post is working with the Trade Desk and other companies to use an email-based identifier for targeted ads). Google argues that, unlike Apple and Mozilla, it actually had small publishers, advertisers and the consumers they serve in mind when it said it would build FLOC to account for the loss of targeting ability when cookies go away. Either way, Google is set up for success. If FLOC does work effectively, it gains more control over the advertising ecosystem and can tell its users it has scored a win for their privacy. If it fails, advertisers will likely invest even more in the "walled gardens" - which conveniently include Google's search ads and YouTube. So what does all this mean for me? The debate over cookies is a major reminder of just how much our online behavior is being tracked and recorded by dozens of private companies. It also shows how many companies have a stake in that reality. Targeted advertising has grown up alongside the Internet, and helped create giants such as Facebook and Google, but also fostered an ecosystem of thousands of companies employing hundreds of thousands of people. When companies such as Google make changes to how products used by billions of people work, there are consequences. Getting rid of cookies completely could hurt news publishers and e-commerce start-ups, decreasing the number of voices online and pushing up prices for consumer products. It could also increase privacy and move the Internet in the direction of less surveillance overall. None of this has been fully decided, and keeping track of the big changes made by companies such as Google, Facebook and Apple over the next several years will be key to understanding how our online lives are recorded, packaged and sold.

GulfNews Technology

How much more of a defence can companies put up against ransomware?

Analysis|Trends|: The trade in cyberattacks is now so advanced that Darkside, the Russia-based group responsible for the Colonial Pipeline attack that shutdown a 5,500-metre-long pipeline in the US, sells its own ransomware software. And even has a tech support service in place should you need additional help in using it. Some cyber-experts have even suggested that this latest attack is essentially a marketing campaign to show just how effective its software is at extorting cash from victims. Colonial were reported to have paid $5 million to take back control of its systems. Not a scattershot approach It has become clear that these attacks are becoming more widespread. In 2020, the UAE recorded a 183 per cent increase in cases where the cyber criminals breach a system and make it impossible for a service to be delivered – known as a Distributed Denial of Service (DDoS) attack. As the Colonial attack has proven, it is no longer small and often defenceless companies that are targets for cyber criminals. Instead of targeting several companies for smaller ransoms, hackers can identify one larger company that has weaker protection or has perhaps neglected its responsibilities towards cyber security. Those millions only help to a point Nevertheless, even organisations that have spent vast sums on cyber security are not immune from attack. We believe there is no ‘one-size-fits-all’ solution to cyber-crime. As it is constantly evolving, ransomware is capable of evading even the most advanced of defences, meaning industry standard security can become obsolete and inadequate very quickly. It is important to have security systems implemented throughout the entirety of a utility network, as hackers often target cloud-based or managed data centres that are remote enough from the grid to be more easily breached. Internal and external communication channels must also be secure. If there are many levels of security, where components of the whole system are divided into separate branches, it will prove much more difficult to break. A major concern regionally is that now our systems are so digitally advanced and reliant on each other, that if one system becomes compromised others will follow. Rail systems, power plants, water treatment plants all use the same technologies, so it is critical that all, not just one of these systems, are fully protected. Most exposed Cybercriminals have taken advantage of the pandemic by attacking at a time when many organisations are at their weakest. Tightened budgetary controls and home working has diverted attention away from IT and info-security concerns, leaving vulnerabilities throughout networks. Despite it being inconvenient to users, multi-authentication practices should be introduced in certain circumstances where sensitive information could be breached. One-time passwords and verification codes are examples. Regular security audits are recommended to help identify areas of susceptibility, too. Outside of a tightened regulatory environment, there are practices that companies can adopt to limit their exposure. Resilient and hard to breach sensors combined with highly secure communication and analytics systems are a strong pillar to the entire security system. Regardless, if resources are not devoted to the problem in sufficient measures, problems will remain. If the pandemic has taught us one valuable lesson, there were warning signs that were ignored, which resulted in catastrophic repercussions. Have we now done enough to be truly confident that our IT systems are safe and secure? Maybe, but what is protected today may not be tomorrow… Francois Frigaux, Special to Gulf News The writer is Director at Sensus, a Xylem brand.

GulfNews Technology

Spread of new skills within organisations cannot be done with a silo mindset

Analysis|Companies|: After all the upheaval we’ve faced in the last year, we hardly need more workplace disruption. But World Economic Forum’s latest Future of Jobs report predicts that machines and AI technologies will take over half of all work tasks by 2025. Where does that leave us ? A pessimist might say that our role in business is coming to an end and that more sophisticated automation and AI had put paid to that. It’s difficult to take that view seriously following the pandemic, though. Where machine learning models struggled to adapt to sudden market changes - and algorithms failed to predict school grades - humans stepped in and saved the day. What businesses needed more than ever was common sense and humanity – that only humans could deliver. True, we’ve only scratched the surface of what AI is capable of. But nothing can beat the adaptability and creativity of human minds. It is thus the HR leaders who are charged with unleashing and harnessing that potential, nurturing it and giving it space in the business to grow. What new skills pay the bills? Every business and industry is changing in its own way, and each requires its own new competencies. In supply chain, the focus is on resiliency and the integration of self-healing capabilities to guard against future shocks. For customer experience, the emphasis is on embedding data-driven customer understanding that creates more tailored ‘human’ experiences digitally. However, despite the supposed chaos, there are strong similarities beneath the surface. Organisations are rapidly embracing an enormous depth of business applications. These cloud apps rely on AI to help people integrate, orchestrate and automate, helping organisations adapt and transform when they need to. We’re at the point where AI not only automates mundane and repetitive tasks, but can also perform highly complex tasks - like business forecasting - accurately and reliably. This doesn’t mean AI is about to replace us – in fact, the inverse is true. Almost every worker, whatever their role, could benefit from the use of a cobot or ‘digital twin’. Resetting workplace culture If many workers are denied access to the latest cloud and AI tools, meaning businesses can’t reap the rewards either. To level up this learning across a business, HR leaders must think differently – but they’ll also need to be more targeted in their approach. They not only need an understanding of what skills are lacking right now, but what talents will be commodities in the future. How does a company know what skills to teach, and which employees are best able to take the business forward? These decisions have to be guided by HR data from all lines of business. Everything from payroll information to messaging and sentiment analysis should be considered to understand where employees may be struggling, and what they need to improve. The larger the organization, the more complex it becomes to maintain a programme of constant learning. That said, during the pandemic, low-carbon energy leader Engie extended its cloud-based HCM platform to maintain business transformation across its global 170,000- strong workforce. By integrating data from across the business and using AI tools, Engie was able to streamline and standardise a global policy of continuous improvement, career development and best practice sharing across the business. Humans make a company’s culture – but helping them learn and grow with the right tech can make it even better. HR plays a crucial role, and when aided by the right AI tools and data-driven insight, they can make the best decisions for the business and its people. No matter what shape the next great disruption takes, HR leaders can lay the groundwork work now to ensure employees can meet the challenge head on. Yazad Dalal The writer is Oracle's HCM Strategy Leader for EMEA territory.

GulfNews Technology

Movie theatres should try and build a working relationship with streaming platforms

Analysis|Trends|: The way we consume content has transformed significantly and is happening at multiple levels. We now consume content on more screens and the amount of content produced has also increased exponentially as new digital mediums make the economics of niche content viable. Currently there are four dominant channels to consume paid content: Movie theaters: This is the one of the oldest modes of content consumption. Consumers generally go there for the immersive experience and pay on a per-view basis. Cable TV: This is largely streaming for sitcoms and live TV driven content. The most common model is the monthly subscription plan. Many of us would have this as they are typically bundled with telephone and internet connectivity. Over-the-top (OTT) bundles: These are on-demand content providers - typically through an app on a smart device. The pricing model is generally monthly subscriptions that allow users to watch all of the content on the platform. Netflix, Amazon Prime and Shahid are some dominant players in this category. Pay per-view (PPV) OTT: This mode is generally for live events and to watch content on a pay-as-you-go basis. Apple’ iTunes, GooglePlay, Zee5 and Mxplayer are some OTT players who have pay-per-view options. Taking them inhouse Recently, the biggest OTT platforms such as Netflix and Amazon have started producing movies for their own platform. In this model, they pay a bulk fee for an already produced movie or produce it themselves. Over-the-top players are now the dominant entertainment channels with consumers spending almost 20 hours per week in Middle East. This is twice the time consumers spent on OTT platforms just two years ago - and many times more than spent in movie theatres. Such large-scale OTT usage presents an opportunity for the movie ecosystem to reinvent and provide more choices to consumers. Link up on OTT For example, new movies can be released on a per-view model on OTT platforms, thus providing more choices to movie studios and also to consumers. It can make economics of niche movies viable and can drive Arabic content production in the region. Even movie theatres can reimagine themselves and create a pay-per-view revenue sharing channel that complements the existing physical movie business. That way they will be able to ride the OTT wave and make it more efficient for all stakeholders in the process. Movie theatres will continue to be relevant for the immersive and social experience they provide. For this to really take off, privacy concerns will need to be addressed. But OTT does open up more possibilities for new movies. We are already seeing in this in some form – it is only a matter of time for it become mainstream. Sandeep Ganediwalla The writer is regional Partner at RedSeer Consulting.

GulfNews Technology

Bank, airline web outage 'not caused' by cyberattack

Technology|World|Business|: Sydney: A major online outage that hit bank and airline websites on both sides of the Pacific was not caused by a cyberattack, the tech provider responsible said Friday. In a statement, US-based Akamai said around 500 of its customers were briefly knocked offline on Thursday because of a problem with one of its online security products. The outage hit American, Delta, United and Southwest Airlines and most of Australia's major banks, leaving customers unable to access websites and mobile apps. also read Internet outages briefly disrupt access to websites, apps Fastly blames software bug for major global internet outage Akamai said the problem was resolved in just over four hours, although most websites experienced problems for around an hour. "The issue was not caused by a system update or a cyberattack," Akamai said, adding the issue had been narrowed down to a data routing problem that had now been fixed. It is the latest incident to draw attention to the stability of economically vital online platforms and the key role that a handful of little-known "CDN" - content delivery network - companies play in keeping the web running. Last week, US media and government websites, including the White House, New York Times, Reddit and Amazon were temporarily hit after a glitch with cloud computing services provider Fastly. Fastly offers a service to speed up loading times for websites. Akamai offers a range of similar IT products designed to boost online performance and security. The firm said this problem was linked to a product that prevents DDoS attacks - an often crude cyberattack that knocks websites out by peppering them with requests for data. "Many of the approximately 500 customers using this service were automatically rerouted, which restored operations within a few minutes," the company said. "The large majority of the remaining customers manually rerouted shortly thereafter."