Technology Archives - The Cyber Shark

Artificial Intelligence Will Reshape the Workforce, Skills, and Communication, Warns LinkedIn CEO

In a recent interview with Bloomberg, LinkedIn CEO Ryan Roslansky discussed how artificial intelligence is changing workplace communication, skill development, and the job market. This week’s interview demonstrates how AI technologies are affecting top executives at Microsoft and other organisations, as well as entry-level job applicants. According to Roslansky, artificial intelligence is both disrupting and facilitating advancement in fields including self-paced skill development, remote learning, and entrepreneurship. As AI transforms traditional roles and work expectations, he emphasised the necessity of ongoing reskilling. “A lot of uncertainty and disruption are going to occur. But in the long term, it’s advantageous,” Roslansky told Bloomberg. As an example of how AI tools are now helping even senior leadership with communication, he acknowledged in a light-hearted moment that he had used Microsoft’s Co-pilot AI to improve the tone of his emails to Microsoft CEO Satya Nadella. LinkedIn trends reveal: Surge in AI-related job listings. More users are adding AI skills to their profiles. LinkedIn is rolling out AI resume builders and job analysis tools. However, users remain cautious. Overuse of AI-generated posts and profiles has affected credibility, prompting LinkedIn to roll out ID and work email verification tools. Roslansky stated that while LinkedIn is implementing an “AI-first mindset” throughout all organisations going forward, the layoffs were not connected to AI. Advice Key Highlights: AI is both disrupting and enabling the job market. In the AI era, flexibility and retraining are crucial. LinkedIn is using AI technologies to improve communication and job readiness. Career Advice in the AI Age: Take tech-focused and AI-focused courses to keep your abilities up to date. Don’t seem robotic when using AI tools, such as resume builders. Verify your LinkedIn profile to boost trust. Keep up with the latest developments in AI that affect your sector. AI should be used as a tool, not a substitute.

Shoplifters are apprehended by this AI technology before they even steal anything

A Paris-based start-up, Veesion, has created an AI surveillance system that can detect suspicious gestures before shoplifting occurs. Launched in 2018, the tech is now used in over 5,000 stores across Europe, the US, and Canada, helping retailers prevent theft before it happens. Without keeping personal information or utilising biometrics, Veesion’s gesture-recognition AI examines body movement to identify questionable behaviour. The solution delivers personnel real-time notifications through a mobile app and connects with current CCTV setups. “The algorithm is indifferent to people’s appearance. As stated by Benoit Koenig, a co-founder of Veesion, “It only cares about how your body parts move over time.” It detects actions like concealing items or irregular handling of products, and immediately alerts the security team with video evidence. Within three months, a shop in the United States reduced losses in the health and beauty department by fifty per cent. The startup recently raised ₹373 crore ($43 million) in Series B funding to expand globally. The tech also helps detect employee theft, which accounts for similar losses as customer shoplifting. “It acts as a deterrent. Staff know AI is watching,” Koenig added. Advice: Veesion’s AI shows how gesture-based surveillance could revolutionise theft prevention in physical stores, without invading privacy. Security Insights for Retailers: Upgrade to AI-enhanced CCTV systems Train staff to respond to real-time alerts Use non-biometric solutions to respect privacy laws Regularly audit security footage for internal theft

16 Billion Logins Stolen In Mega Data Breach Threatening Apple, Google and More

In a massive global data breach, hackers have leaked 16 billion login credentials, exposing users of platforms like Apple, Google, Facebook, GitHub, and even government services, according to Forbes. The breach was uncovered this week by cybersecurity researchers who found the records in 30 unprotected datasets, some containing up to 3.5 billion entries each. Cybersecurity experts first noticed a mysterious open database containing 184 million records, which has now proven to be only a small part of a much larger breach. Researchers have since confirmed 30+ datasets leaked online, containing VPN logins, social media accounts, and developer platform access. The breach is not a recycling of old data but fresh, high-value information collected throughout 2025, making it dangerous and exploitable for phishing attacks, account takeovers, and business email compromise (BEC) schemes. “This is not just a leak — it’s a blueprint for mass exploitation,” warned the research team cited in Forbes. Darren Guccie, CEO of Keeper Security, emphasised that the leak’s scale and nature pose “far-reaching risks” to users across sectors. Google has responded by urging users to move beyond traditional passwords and 2FA, and instead adopt passkeys, which are biometric login systems using fingerprint or facial recognition. “Passkeys are phishing-resistant and safer than passwords,” Google stated. Advice: This mega breach highlights how credential leaks remain one of the top cyber threats globally. The fact that the stolen logins are fresh and tied to major platforms increases the urgency for users to act. Cyber Safety Tips: Switch to Passkeys or biometric login methods wherever possible Update your passwords immediately, especially for sensitive accounts Enable multi-factor authentication (MFA) Avoid clicking on suspicious links or emails asking for logins Report scams or suspicious activity at cybercrime.gov.in

Google rolls out budget-friendly Gemini 2.5 Flash Lite, opens 2.5 Flash and Pro to all

New model offers high-speed, low-cost performance; Gemini 2.5 Flash and Pro now open to all users. Google has launched a new AI model, Gemini 2.5 Flash-Lite, on June 18, 2025, making it the fastest and most affordable in the Gemini 2.5 series. Available via Google AI Studio and Vertex AI, it supports multimodal tasks with low latency and cost, catering to developers and enterprise users who require scalable AI solutions. The Gemini 2.5 Flash-Lite is designed for high-volume, latency-sensitive tasks such as translation, classification, and reasoning, offering improved performance over its predecessor, the 2.0 Flash-Lite. Google claims it delivers superior accuracy in coding, science, and multimodal benchmarks while being cost-efficient. Despite being a “lite” version, it includes advanced features such as a 1 million-token context window, tool integration (like Google Search and code execution), and flexible compute scaling based on budget. Google has also announced the general availability of Gemini 2.5 Flash and Pro, previously limited to select users. Firms like Snap and SmartBear have already integrated them into their production systems with success. These models are now accessible via Google AI Studio, Vertex AI, and the Gemini app, expanding usage beyond developers to general users through tools like Search. Quote: “Gemini 2.5 Flash-Lite is designed to bring scalable, affordable, and high-performance AI to everyone — from individual developers to large enterprises,” said a Google spokesperson in the launch statement. Advice: Google’s move to open Gemini 2.5 models, especially the new Flash-Lite, offers a powerful AI toolkit for cost-conscious developers and businesses. Its lightweight design doesn’t compromise capability, making it ideal for fast, real-world deployment. Tips for AI users: Choose models based on task latency and cost requirements Explore Google AI Studio or Vertex AI for hands-on testing Use Flash-Lite for rapid classification, translation, and large prompt processing Review Google’s documentation to integrate AI efficiently and securely

Google Launches AI Safety Charter to Combat Rising Cybercrime Threats in India

With 60 million threats blocked and ₹5 million pledged, Google partners with Indian agencies and IIT-Madras to secure the digital future In India’s rapidly growing digital space, hackers are increasingly targeting users with sophisticated scams, including fake banking apps and phishing attacks. This week, Google revealed new efforts under its AI-led “Safety Charter” to detect and stop such scams in real-time through artificial intelligence. The goal: to protect Indian users and institutions from cyber fraud and data theft. Google is leveraging AI to block scams before they impact users. According to Preeti Lobana, Google India’s Country Manager, the company has already: Prevented 60 million attempts to install malicious applications Flagged 500 million scam messages Marked 12 million fake listings on Google Maps Issued 2.5 billion warnings for suspicious URLs Financial frauds now account for 75% of cybercrimes, with global cybercrime costs estimated to reach $14 trillion by 2028. To tackle this, Google has introduced a three-pronged AI Charter focused on: User safety Enterprise and government cybersecurity Responsible and ethical AI development In addition to deploying advanced AI tools, Google is collaborating with Indian authorities such as the Department of Telecommunications, the Home Ministry, and SEBI for centralised intelligence sharing. Google.org also announced a $5 million expansion of the APAC Cybersecurity Fund via The Asia Foundation, enabling cyber clinics and training in Indian universities, especially for MSMEs and students. Meanwhile, a new partnership with IIT-Madras aims to push forward Post-Quantum Cryptography, securing future digital interactions through anonymous, privacy-first tokens. According to Lobana, “AI systems are always evolving to detect new threats—even those that haven’t been seen before.” Additionally, Heather Adkins, Google Security’s VP of Engineering, said: “AI isn’t just narrowing the gap between attackers and defenders—it’s eliminating it in some cases.” Advice: Google’s AI-driven defence strategies signal a major leap in protecting India’s digital users. But while tech giants act, user vigilance remains essential. Security Tips for Users: Turn on two-factor authentication (2FA) for every application. Never click on unknown links in emails or communications. Download apps only from official stores Report cybercrimes at cybercrime.gov.in Stay updated with Google’s Safety Centre and CERT-In alerts

Huge discovery: DRDO-IIT Delhi milestone reshapes future warfare and guarantees India’s leadership in quantum technology

India pioneers quantum-secure free-space communication with DRDO-IIT Delhi breakthrough, marking a leap in defence, cybersecurity, and future warfare capabilities. With the successful demonstration of quantum entanglement-based free-space secure communication, carried out in collaboration with the Defence Research and Development Organisation (DRDO) and IIT Delhi, India has made a significant step toward the quantum era. Secure quantum communication across a distance of more than one kilometre was accomplished in the experiment, which was conducted at the DRDO-Industry-Academia Centre of Excellence (DIA-CoE) on the campus of IIT Delhi. With the use of a free-space optical link, the milestone achieved a safe key rate of around 240 bits per second and a quantum bit error rate of less than 7%. This lays the groundwork for long-distance Quantum Key Distribution (QKD), real-time applications in quantum cybersecurity, and the larger goal of creating quantum networks and a future quantum internet. The Directorate of Futuristic Technology Management (DFTM), DRDO, has approved the project named “Design and development of photonic technologies for free space QKD.” DRDO’s DG (MED, COS & CS), Director SAG, Director DFTM, and top academics from IIT Delhi were among the senior authorities who attended the demonstration, which was headed by Prof. Bhaskar Kanseri’s research group. Higher security is guaranteed by quantum entanglement-based communication, in contrast to conventional prepare-and-measure QKD systems, even if the hardware is compromised. The entangled quantum state is changed by every attempt at eavesdropping, making it observable. Additionally, it avoids the need to install expensive and disruptive optical fibre networks, which is especially advantageous in urban or remote locations. This accomplishment builds on earlier successes, such as the 2022 establishment of India’s first intercity quantum communication link between Vindhyachal and Prayagraj and the 2024 demonstration of a 100-km fibre-based quantum key distribution system. DRDO’s DIA-CoEs, which currently encompass 15 prestigious academic institutions nationwide, serve as the foundation for all initiatives. “India has entered into a new quantum era of secure communication, which will be a game changer in future warfare,” said Defence Minister Rajnath Singh, congratulating the team. Both IIT Delhi Director Prof. Rangan Banerjee and DRDO Chairman Dr. Samir V. Kamat reiterated the view, praising the discovery as a crucial step towards technical sovereignty and national security.

Facebook clean-up: Why Meta removed over 23,000 accounts in Brazil and India

Facebook clean-up: Meta removed over 23,000 Facebook accounts in Brazil and India linked to scams using deepfakes of public figures to promote fake investments and gambling apps. The company is enhancing security features and collaborating with Indian authorities to combat online fraud and protect users. On 07-05-2025, Meta announced that it had deleted over 23,000 Facebook Pages and accounts that were part of extensive scam operations that mostly targeted people in Brazil and India. The March takedown was a component of Meta’s larger initiatives to fight online fraud and increase public awareness of digital fraud. Meta claims that the fraudulent networks used deep fake technology to pose as well-known individuals, such as corporate leaders, cricket players, and producers of financial information. These edited movies purportedly featured public individuals promoting gambling apps and fraudulent investment schemes. Under the pretence of getting investment advice, consumers were taken to messaging platforms after interacting with the content. In several cases, consumers were tricked into downloading harmful gambling apps from a fake website that looked like the Google Play Store. Meta outlines several common scams. Meta described some increasingly common scam types as part of its effort to improve user protection. While payment scams usually take advantage of internet anonymity and urgency, investment scams frequently offer large returns on assets like shares, real estate, or cryptocurrency. Scammers who pose as vendors on websites like Facebook Marketplace and demand cash before vanishing without delivering the promised products are known to engage in advance payment fraud. Other tactics include Over Payment Frauds, in which con artists request a refund after claiming to have overpaid for an item, only to later reverse their initial (often fraudulent) transaction. As a result, Meta has implemented some security features on Facebook, Instagram, and WhatsApp, including scam alerts in Messenger and instructional materials to assist users in identifying fraudulent activities. To improve online safety, the internet giant said it is collaborating closely with Indian government agencies and civil society groups. Initiatives mentioned include cooperation with the Indian Cybercrime Coordination Centre, the Department of Telecommunications, and the Department of Consumer Affairs. Reiterating its dedication to digital security, Meta urged customers to stay alert and make use of the resources available to report and steer clear of frauds.

YouTube pilots two-person Premium plan to boost its subscription push

YouTube is piloting a two-person Premium plan in India and select regions to offer users a more affordable subscription option. Priced at ₹219/month in India, the plan targets couples or cohabitants and supports YouTube’s broader push to grow subscription revenue. YouTube is testing a new subscription tier that allows users in Taiwan, Hong Kong, France, and India to share their Premium or Music Premium membership with a household member. This action coincides with the Google-owned video platform’s increasing emphasis on subscriptions as a possible source of revenue as part of its strategy to expand beyond advertising. The YouTube Music Premium subscription tier is accessible for Rs 149 per month during the pilot. In contrast, the two-person YouTube Premium subscription tier is being provided in India for Rs 219 per month. Participants must be in the same Google family group, have a Google account, and be at least 13 years old. The Cyber Shark verifies this development from YouTube. In an announcement, a YouTube representative stated, “We’re experimenting with new ways to provide greater flexibility and value to our YouTube Premium subscribers, including offering a two-person Premium plan option, allowing two people to share a subscription at a reduced cost.” In addition to the ability to save videos for offline playing and play them in the background, YouTube Premium offers its customers an ad-free experience. Additionally, it gives them early access to certain of the platform’s experimental features. Similar capabilities are available with Music Premium, but just for music content. Monthly rates for YouTube Premium in India start at Rs 89 for students, Rs 149 for individuals, and Rs 299 for families. Students’ Music Premium plans start at Rs 59, individuals’ plans at Rs 119, and families’ plans at Rs 179. Couples, roommates, or any two people living together who want to lower subscription expenses while keeping separate accounts will probably find the two-member plan useful. Spotify, a massive music streaming service, also provides its users with a comparable package called Duo. Google has increased the cost of its premium subscription in numerous areas and launched a global campaign against ad blockers as part of its efforts to boost YouTube’s subscription income in recent years. Depending on the tier, price increases in India in August 2024 ranged from 12% to 58%. “A significant portion of YouTube’s revenue comes from subscriptions. During the business’s April 2025 earnings call, Sundar Pichai, CEO of Alphabet, the parent company of Google, stated, “We are continuing to expand our subscription options. We recently expanded our Premium Lite pilot to the US, giving users a new way to enjoy most videos on YouTube ad-free.” Viewers will be able to watch ad-free videos in a variety of genres, including learning, comedy, gaming, and culinary, with Premium Lite. However, they will see advertisements for shorts, music videos, and music content. Additionally, customers won’t be able to download videos for offline viewing or play them in the background. In addition to the US, the plan is presently offered in Thailand, Germany, and Australia. YouTube reported in March that it has more than 125 million subscribers worldwide across its Music and Premium services. During the April earnings call, Pichai stated that YouTube and Google One were the main factors behind Alphabet’s overall success in surpassing 270 million paid memberships. Google usually only provides the platform’s advertising sales numbers each quarter, not the total amount of money made by YouTube. But in October 2024, Pichai said that for the first time in the previous four quarters, YouTube’s total revenue had topped $50 billion.

Pinterest is at last taking something to combat its AI infestation

Pinterest has introduced new tools to label and filter AI-generated images, helping users distinguish synthetic content and reduce its visibility. The platform will use metadata and AI classifiers to detect altered images, with features like “AI modified” tags and user controls to limit such content. On its site, Pinterest is helping users recognize and steer clear of AI-generated slop. To help customers avoid seeing too many images while searching for related topics, the business is introducing new capabilities that will automatically classify images that are identified as being created or altered using generative AI. “As our users encounter AI-generated content on Pinterest, we are helping them make better choices about what they see,” said Chief Technology Officer Matt Madrigal. “We are consciously tackling this new terrain in a way that is advantageous to all users on Pinterest, and Gen AI content on Pinterest ought to improve users’ capacity to find and act on their inspiration.” Global rollout of Pinterest new Gen AI labelling feature should help shield users from fraud. When a pin is clicked on in close-up, the labels will show up as a “AI modified” stamp in the lower left-hand corner. Pinterest uses metadata analysis, most likely for invisible marks like Adobe’s Content Credentials or Google’s Synthid, to determine whether an image was created or altered using artificial intelligence. Additionally, Pinterest claims to have created “classifiers that automatically detect gen AI content,” even in cases where an image lacks metadata markers. Although detection-based AI flaggers are not always accurate, Pinterest is now enabling users to file an appeal if they think their pins have been incorrectly labelled. It would be an understatement to say I’m overjoyed. The dominance of AI photos on Pinterest has drawn a lot of criticism from users, who find it hard to use the mood board builder for anything useful other than putting together pictures for aesthetic reasons. For instance, it might be difficult for artists to locate reliable real-world reference materials, and, likely, the apparel, accessories, or furnishings items seen in pins aren’t available for purchase because they don’t exist. Even as a source of “inspiration,” that can be problematic because artificial intelligence (AI) can produce incredibly realistic-looking hairstyles and interior design concepts that may be impossible to accomplish in real life. Additionally, an experimental tool will be released “soon” that will let users filter out specific AI photos for categories like beauty and art that are “prone to AI modification or generation,” according to Pinterest. The three-dot menu located at the bottom right of a pin will have the option to “see fewer.” Although it’s unknown how much AI the feature will filter out, Pinterest claims that this will alert its systems to promote less of such content and will eventually spread to more pin categories. I can only hope that an “all of it” option will be added later.

Google Notebook LM AI podcast expands to more languages, including Hindi

Google has expanded its Notebook LM AI podcast feature, Audio Overviews, to support over 50 languages, including Hindi, Marathi, Urdu, Maithili, and Konkani. The tool uses Gemini’s multimodal AI to convert complex texts into podcast-style conversations, making learning more accessible. More languages are now supported by Google’s Audio Overviews function in Notebook LM, the AI-powered note-taking and research tool. More than 50 other languages are currently supported by the feature, including major Indian languages like Hindi, Maithili, Konkani, Urdu, and Marathi. According to the business, this multilingual capability is made possible by multimodal Gemini’s inbuilt audio support. The change was revealed in a blog post by the US-based software corporation. The whole public can now use the functionality. Overviews of Google Notebook LM Audio: How it Operates Imagine a scenario in which a user comes across a complicated paper or article and finds it challenging to understand. In this scenario, the user can utilise the Audio Overviews capability and submit the content to Google Notebook LM. The uploaded content will be converted by the AI into a dialogue between two AI-generated hosts, like a podcast. Rather than reading the material out loud, the hosts engage the audience in conversation while breaking down complex ideas as needed. This method makes learning more approachable and helps to simplify complex information. The resulting audio transformed a four-minute read into a nine-minute podcast when Business Standard tested the function with one of its articles on direct-to-mobile (D2M) phones. The audio sounded human because the two AI hosts spoke naturally, pausing and stuttering realistically. In order to customise the podcast to their requirements, customers could also designate the topic of the discussion. Audio Overviews are now accessible in over 50 languages thanks to this upgrade. Google first expanded Notebook LM to more than 200 countries last year. Overviews of Google Notebook LM Audio: How to Use Go to the Google Notebook LM official website. Select “Create New.” You can use Google Drive to upload your work by copying and pasting text or by pasting a link. Select “Generate” from the “Audio Overview” menu. Listen to the podcast that the AI created using your content.

Technology