IT News Review by Control F5 Software: Google can share your text messages with your employer on company-managed devices

Google can share your text messages with your employer on company-managed devices

A recent article sparked controversy after explaining how an Android update allows employers to see SMS and RCS messages sent from phones managed through Google Workspace. We are not talking about personal phones, but about company-managed devices, where IT policies may include archiving and monitoring communications, including those sent via Google Messages.

For professionals and IT teams, the novelty underscores the need for clear BYOD policies and transparent communication to employees about what is and is not private on a work device. From a security and compliance standpoint, the functionality helps regulated organizations (financial, legal, governmental), but raises serious trust questions and necessitates strict separation between personal and corporate phones.

Sora and Nano Banana Pro face usage limitations due to huge demand

Google and OpenAI have reduced the number of free generations for their popular image and video models, Nano Banana Pro and Sora, citing very high demand and pressure on GPU infrastructure. Sora users without a subscription can now generate only a few videos per day, while Google limits Nano Banana Pro users to a smaller number of images, noting that these caps can be adjusted at any time.

For professionals using these models in production or prototyping, the change means they can no longer rely on heavy use of the free versions and must plan budgets for credits or dedicated subscriptions. From the software and AI industry perspective, the message is clear: top generative models are rapidly becoming commercial products, with controlled and prioritized access for paying customers, which influences how workflows and infrastructure costs are designed in companies.

The EU checks if Apple Ads and Apple Maps fall under DMA rules

The European Commission has announced that it will analyze whether Apple Ads and Apple Maps should be classified as "essential platform services" under the Digital Markets Act (DMA), which would subject them to stricter obligations. Apple argues that these products do not have the role or position of a "gatekeeper" and that extending the rules would be disproportionate.

For the digital ecosystem, a potential reclassification would mean new rules regarding competitor access, interoperability, and how Apple can integrate services into iOS and the App Store. For businesses that rely on advertising and location services, a more DMA-constrained Apple could mean more transparent conditions, but also possible adjustments to APIs, pricing, and development priorities on the company's side.

Google clarifies how Gmail uses data for "smart" features and Gemini

Google has published clarifications about how the smart features in Gmail, such as Smart Reply, Smart Compose, and the new Gemini for Workspace, process the content of emails. The company insists that user data is not used to personalize ads and that the AI models used in enterprise products comply with existing privacy policies.

For teams working with sensitive data, the message is that adopting these AI features should not change the risk profile compared to "classic" Gmail, as long as administrative settings and internal policies are respected. At the same time, the clarifications show that software producers need to be increasingly explanatory regarding the data pathways behind "smart" features, especially when combining generative AI with confidential company information.

Thousands of Amazon employees call for clearer AI rules in an open letter

Thousands of Amazon employees have signed an open letter addressed to CEO Andy Jassy, calling for a stronger voice for workers in decisions related to the development and use of AI. They describe themselves as "workers who develop, train, and use AI" and demand concrete measures to ensure that the technology is not used abusively or contrary to declared ethical values.

The letter reflects an increasingly visible tension in large tech companies: the pressure to quickly launch AI products versus concerns about their social, ethical, and environmental impact. For technology leaders and product teams, this type of internal initiative shows that AI governance can no longer be treated merely as a PR topic, but must be integrated into real development processes, with authentic feedback from the people who build and operate the systems.

Concerns about employee health at Amazon data centers in Oregon

An investigative report has gathered testimonies from current and former employees of Amazon's data centers in Oregon, as well as from local residents, who report serious health issues, including cases of cancer and pregnancy loss. Meanwhile, information has emerged about incidents involving leaks of chemicals used for cooling and maintenance, while Amazon rejects the idea of a causal link and claims to comply with regulations.

For the digital infrastructure sector, the story serves as a reminder that data centers are not just abstract entities in the cloud, but physical facilities with a very concrete impact on local communities and the people who work there. As demand for AI grows and data center parks expand, standards for transparency, safety, and reporting become a critical part of how companies build and operate large-scale infrastructure.

Aqara launches smart home hub with integrated Siri in the wall

Aqara has introduced the S1 Plus, a wall-mounted smart home hub that combines local control for its own ecosystem with integration with Apple Home and Siri commands directly from the device. The hub also functions as an interface for the Matter standard, allowing connection and control of multiple smart home devices from different ecosystems.

For developers and integrators of smart home solutions, this type of hybrid hub strengthens the trend towards interoperability and local control, with reduced latency. From the perspective of advanced users, the ability to run Siri and local automations on a dedicated panel, not just on phones or speakers, opens new integration scenarios with automation, security, and energy management applications.

Intel targets Apple again with a low-end rival to Apple M chips

Intel is developing a new low-end processor explicitly designed to compete with Apple's entry-level chips and to gradually regain some of the lost Mac business. The new design emphasizes energy efficiency and integrated graphics powerful enough for everyday tasks, considering that many Mac users do not fully exploit the performance of the high-end series.

For Windows laptop manufacturers and companies standardizing on certain hardware platforms, a more competitive Intel in the efficiency segment could mean real alternatives to the Apple ecosystem for users such as developers, content creators, and business personnel. At the same time, it puts pressure on software providers to better optimize applications for hybrid architectures, both ARM and x86, especially in client-level AI scenarios.

Google Maps receives a power-saving mode and new features in Pixel Feature Drop

A new "feature drop" for Pixel phones brings a power-saving mode in Google Maps, which reduces resource consumption during navigation by limiting position updates and certain animations. The update is part of a larger package of news for Pixel 10, aimed at better battery life and AI-based features.

For users, this mode is useful in long road scenarios or in areas with difficult access to charging, but also for companies managing mobile fleets and being mindful of battery life in the field. From a software perspective, introducing a dedicated mode for such a widely used application shows how energy optimizations are becoming a standard part of product design, not just a detail at the operating system level.

iPhone 17 and Galaxy S25 dominate smartphone sales in 2025

A recent report shows that the iPhone 17 and Samsung Galaxy S25 are the top models in smartphone sales in 2025, supported by upgrades in cameras, battery life, and system-level AI features. The high-end segment remains the profit engine, even though the overall smartphone market is relatively stable and without spectacular growth.

For mobile app developers and digital service providers, this Apple-Samsung balance in the premium segment confirms which platforms deserve prioritized performance optimizations, local AI integration, and testing of new features. At the same time, the increasing emphasis on on-device AI capabilities means that hardware choice becomes directly linked to how you can deliver intelligent experiences without overly relying on the cloud.

ChatGPT and Copilot leave WhatsApp due to new Meta terms

OpenAI and Microsoft have confirmed that they will stop distributing the ChatGPT and Copilot chatbots through WhatsApp, due to new terms of use that prohibit the use of the WhatsApp Business API for AI chatbots not developed by Meta. The services will remain available until mid-January 2026, after which users will need to access these models through other channels.

For businesses that have built customer support or internal consulting flows on these integrations, the change means a rapid migration to alternatives: proprietary web channels, mobile apps, or other chat platforms. At the industry level, the decision shows how major players in communication platforms are trying to protect their own AI products, which can lead to even greater fragmentation of access points to generative models.

Copilot becomes an AI "personal shopper" in the Edge browser

Microsoft has announced a set of shopping features for Copilot in Edge, aimed at users in the US: price comparison, price history, cashback, price tracking, and "product insights." Copilot can even proactively signal if it finds a better deal for a product you are analyzing.

For merchants and ecommerce platforms, such AI tools embedded directly in the browser change how users discover and evaluate products, shifting some power to AI assistants that filter offers. For B2B solution developers, it becomes increasingly important to integrate structured catalog data and price feeds in ways compatible with these AI agents, either through schema markup or dedicated APIs.

Perplexity launches an AI shopping assistant with PayPal integration

Perplexity introduces a free AI shopping service for users in the US, where you can describe what you are looking for, and the model recommends products in the form of cards with specifications and reviews. Purchases can be made directly from the interface, through an "Instant Buy" integration with PayPal, while still maintaining the relationship between the customer and the merchant, similar to a purchase on the merchant's own site.

This "shopping agent" model in the cloud shifts part of the research and decision-making stage from traditional search engines to conversational assistants. For businesses selling online, it means that optimizing presence in front of these agents (clean data, clear descriptions, authentic reviews) becomes as important as optimizing for SEO or traditional marketplaces.

OpenAI confirms the first AI hardware prototype created with Jony Ive

Sam Altman and designer Jony Ive have stated that they are working on a hardware device prototype for OpenAI, with a design considered mature enough to advance to production in less than two years. The device is described as being about the size of a smartphone, likely without a classic screen, with a "simple, beautiful and playful" design.

For the industry, a hardware product signed by OpenAI would mark the transition from AI as a pure cloud service to embedded AI in dedicated devices, with natural interaction and permanent access to models. For companies building applications over the OpenAI ecosystem, this type of device raises questions about new integration points, voice or gesture-based UX, and scenarios where AI becomes a physical companion instead of just an application.

Anduril faces issues in tests with autonomous weapons

A Wall Street Journal report, picked up by TechCrunch, shows that the defense startup Anduril has encountered numerous problems in tests with autonomous systems: naval drones that failed during a US Navy exercise, mechanical incidents with the pilotless Fury aircraft, and a test of the Anvil system that caused a 22-acre fire. In Ukraine, the loitering drones Altius also reportedly had a high failure rate.

Anduril claims that such difficulties are normal in weapon development and that engineers are making progress, but the situation highlights the current limits of applying AI and autonomy in combat environments with very high stakes. For the civilian tech sector, the story is a reminder that the transition from spectacular demos to robust performance in the field is a long process, with many iterations and operational risks.

xAI builds a small solar park near the Colossus data center

Elon Musk's xAI company plans to build a small solar park near the Colossus data center, aimed at powering part of the energy needs for training and running AI models. The project is presented as a step towards reducing dependence on traditional grids and the carbon footprint associated with AI infrastructure.

From an infrastructure perspective, the initiative underscores the enormous pressure of AI on electrical grids and the growing interest in dedicated sources of renewable energy built alongside data centers. For cloud operators and software providers, such projects can become competitive differentiators, especially for enterprise clients sensitive to ESG criteria and long-term energy stability.

YouTube is working on a mode that "curates" the Home feed

YouTube is testing a feature that allows users to more clearly separate types of content on the main page, reducing the confusing mix of Shorts, long videos, and old recommendations. The new option would provide more control over the feed, possibly through filters or dedicated tabs, so users can more easily see what is currently relevant.

For creators and brands, such a change could affect how content reaches the audience organically and may make the format strategy (Shorts vs. long video) even more important. From the perspective of digital product developers, it is yet another example of a platform trying to fix the overlap between recommendation algorithms, multiple formats, and user fatigue.

Character.ai launches interactive stories for children

Character.ai is preparing a child-oriented product based on guided interactive stories, instead of completely open chat with AI personas. The idea is for young children to interact with characters and narratives designed specifically for their age, with clear content limits and a higher degree of control from parents.

For the AI industry, this more structured format is an attempt to combine the benefits of conversational interaction with the strict safety and compliance requirements for minors. For companies building educational or entertainment products for children, the example shows a realistic direction: AI agents with predefined, auditable scenarios, instead of completely open models.

Speechify adds dictation and voice assistant in the Chrome extension

Speechify has expanded its Chrome extension with voice typing features and a small voice assistant that can take commands to read, summarize, or manage content on the page. Users can dictate directly into text fields in the browser, and the extension converts voice to text, integrating with already used web applications.

For professionals, the combination of text reading and dictation reduces friction in working with documents, emails, and online forms, especially in repetitive workflows. For SaaS product developers, the fact that such features come at the browser level means that part of the "magic" shifts from the application to cross-sectional extensions, which can influence roadmap and integration decisions.

Google and Accel seek the next generation of AI startups in India

Google has partnered with the investment fund Accel for a program dedicated to AI startups in India, focusing on foundational models and enterprise applications. The initiative combines early-stage funding, access to Google Cloud, and technical mentoring for selected teams.

India is already an important hub for software development, and such a program could accelerate the emergence of local companies building globally competitive AI products. For the global software ecosystem, it also means more solutions built natively on Google Cloud, with direct integration into its AI stack, from large models to observability and security tools.

Anthropic launches Opus 4.5 with integration in Chrome and Excel

Anthropic has presented the Opus 4.5 model, an upgrade to its line of top models, focusing on better reasoning and lower response time. The interesting novelty for business users is the appearance of dedicated integrations, including an add-on for Excel and an extension for Chrome, which bring the model directly into daily work tools.

For software teams and analysts, this means they can use an advanced model for data analysis, formula generation, or information cleaning without leaving spreadsheets or web applications. From the solution architecture perspective, these integrations mean less "context switching" and a step forward towards agents that connect directly to production data, with all the security and governance implications that entails.

MEPs want a minimum age of 16 for access to social networks

The European Parliament has adopted a political position calling for the minimum age for access to social platforms to be raised to 16 in the EU, with exceptions for educational services and other specific contexts. The proposal also includes stricter obligations for platforms regarding the protection of minors, anti-addiction design, and age verification.

For software companies and social platforms, such a change would involve serious investments in robust age verification mechanisms, as well as changes in the design of recommendation algorithms and interfaces. For businesses marketing to mixed-age segments, it becomes crucial to reposition channels and find ways to communicate that comply with the new rules without compromising commercial objectives.

OpenAI estimates up to 220 million ChatGPT paying users by 2030

According to an internal note reported in the press, OpenAI estimates that it could reach up to 220 million paying ChatGPT users by 2030, as enterprise products and B2B integrations mature. The projections go hand in hand with plans to expand infrastructure and partnerships with major cloud providers.

For the market, such figures suggest that conversational AI is seen as a standard component in companies' digital stack, similar to email or productivity suites. For developers and software providers, it implies a greater dependence on central AI platforms and, implicitly, the need for clear multi-cloud strategies, portability, and long-term cost negotiations.

Warner Music Group signs a licensing agreement with Suno AI

Warner Music Group has concluded an agreement with Suno AI, one of the generative music startups, to license the catalog and establish rules for the use of content in generating new songs. The agreement includes both the training of models and how the generated results can be marketed.

For the music industry and AI startups, it is an important step towards legitimate business models, where training on the catalogs of major labels is done with a license, not in the gray area. For software platforms that will integrate AI-generated music into games, applications, or campaigns, such agreements pave the way for clearer commercial uses, with less legal risk.

A former content strategist from MrBeast builds an AI tool for creators

A former content strategist from MrBeast's team is developing an AI product dedicated to creators, focused on generating content ideas, testing titles and thumbnails, and performance analysis. The tool promises to combine historical engagement data with generative models to provide concrete suggestions before a clip is produced and published.

For content creators and brands that seriously invest in video, such a tool can shorten the experimentation phase and improve the success rate of each material. From the perspective of software teams, it is a clear example of a product built "by creators, for creators," using AI not just for generation but also for analytics and decision optimization.

European agreement on payment services: more protection against online fraud

EU negotiators have reached an agreement on a new legislative package for payment services aimed at reducing online fraud and hidden fees. Among the measures are stricter authentication obligations, increased transparency on costs, and clearer rules regarding provider liability in case of fraud.

For fintechs, banks, and payment providers, this means adjustments to UX (especially in authentication flows) and additional investments in fraud detection and customer notification channels. For developers building ecommerce applications or digital wallets, the new framework raises the bar in terms of security and audit, but can enhance user trust in online payments.

Baidu becomes a key player in China's AI chip market, replacing Nvidia

Amid US restrictions on GPU exports to China, Baidu has accelerated the development of its own AI chips and is becoming an important supplier for local clients, including large companies and public institutions. Its chips are used for training and running generative models, covering part of the gap left by Nvidia.

For the global software and AI ecosystem, this signals the emergence of an increasingly fragmented supply chain, where hardware and software solutions are closely tied to geopolitical blocks. For technical teams, it means that the portability of models and MLOps systems between different architectures (Nvidia, Baidu, Huawei, etc.) will become an important design criterion.

Europe bets on a slow but steady pace in the AI race

The "slow and steady" approach of Europe in the AI field, focusing on regulation and data protection, could become a competitive advantage over time. Instead of just chasing the speed of product launches, the EU is investing in frameworks like the AI Act and in common data infrastructures to ensure interoperability and trust.

For software companies and AI service providers operating in Europe, this means more bureaucracy in the short term, but also a more predictable and credible environment for enterprise clients and the public sector. As large clients increasingly emphasize governance, audit, and explainability, a robust framework could turn the current "slowing down" into a concrete commercial advantage.

Google Cloud and AWS jointly launch a fast multicloud connectivity service

Google Cloud and AWS have announced a joint service that allows customers to create high-speed private connections between their infrastructures in both clouds. The solution combines AWS Interconnect with Google’s Cross-Cloud Interconnect and introduces an open specification for network interoperability that can also be adopted by other providers.

For solution architects and DevOps teams, this type of dedicated connectivity reduces reliance on public internet for traffic between clouds and simplifies the implementation of active-active multi-cloud architectures. At the same time, it remains to be seen how traffic costs between clouds will be structured and whether the technical benefits will offset the additional expenses for medium-sized companies.

Unpowered SSDs can gradually lose data over time

A technical article from XDA explains why SSDs used as "drawer" storage media, without power for years, can start to lose data. NAND memory retains information in the form of electric charge, and in the absence of power, this charge gradually degrades, especially at higher temperatures and on denser generations of NAND.

For IT professionals and teams managing backups, the conclusion is that SSDs are not the safest option for long-term archiving without periodic checks. Classic solutions like HDDs, magnetic tape, or specialized optical media remain more suitable for "cold" archives, while SSDs are ideal for active storage and caching.

China warns of a possible "bubble" in humanoid robots

The national economic planning agency of China has warned that a possible bubble is forming around humanoid robots, with over 150 active companies and large investments, despite still having few proven commercial users. Officials speak of the risk of very similar models emerging, while funding for real research begins to dilute.

For the robotics industry and investors, the message is that hype cannot replace product-market fit, and real adoption will depend on concrete use cases, not just spectacular demonstrations. From a software perspective, the concept of "embodied intelligence" remains promising, but success will depend on integration with robust AI systems, sustainable costs, and scalable maintenance models.

A new AI benchmark tests how well chatbots protect human well-being

Building Humane Technology has launched HumaneBench, a benchmark that evaluates not only the intelligence of AI chatbots but also how they protect the psychological well-being of users. The benchmark uses hundreds of realistic scenarios, from mental health issues to technology addiction, to see when models "fail" under the pressure of instructions that ask them to ignore humane principles.

The results show that many models become harmful when explicitly asked not to prioritize user well-being, while some perform better when trained to respect clear human-centric design principles. For companies building products over large models, such benchmarks raise the question: what does "safety by design" mean in concrete terms, and how can it be systematically tested, not just through internal policy.

Nvidia CEO Jensen Huang wants any possible task to be automated with AI

A Fortune article quotes an internal meeting where Jensen Huang reportedly told Nvidia managers that it is "madness" to discourage the use of AI and that "any task that can be automated with artificial intelligence should be automated." At the same time, he promised employees that they would not be left without work, as AI will generate new types of tasks.

The message combines an aggressive "AI-first" philosophy with a promise of reassurance for the workforce, reflecting how major hardware and software providers are trying to lead the technological transition. For companies adopting AI, the statement reinforces the idea that it is not just about "adding AI" to a product, but about rethinking entire business and software development processes around intelligent automation.

Synthesis made with the help of a monitoring flow provided by Control F5 Software.