IT News Review by Control F5 Software: A satellite has independently identified objects in orbit with the help of AI

A satellite has autonomously identified objects in orbit using AI

An Earth observation satellite has successfully autonomously identified what it was looking for for the first time, without human analysis on the ground. The demonstration took place in April 2026 on YAM-9, a spacecraft built by Loft Orbital, using NASA JPL software and the Gemma 3 model from Google DeepMind.

The system used an in-orbit vision-language model to analyze images and respond to requests in natural language. Instead of the satellite sending large volumes of data to the ground, processing took place directly on board.

For the space industry, this could change the value of orbital sensors. If satellites can locally identify objects, infrastructure, or events, relevant data can be sent more quickly, and analysis costs can decrease.

For AI applied at the edge, the experiment is significant. It shows that multimodal models can function in resource-limited environments, where latency, connectivity, and autonomy are critical.

Social media becomes the main source of news

Social media has surpassed traditional sources of information globally for the first time, according to the Digital News Report 2026 conducted by the Reuters Institute. The study shows that 54% of respondents used social media as a news source at least once a week, compared to 51% who used television, radio, or news websites.

This change confirms an important transition in how the public discovers, consumes, and validates information. Social and video platforms are no longer just distribution channels; they are becoming primary information spaces, influenced by algorithmic recommendations, independent creators, and fragmented consumption behaviors.

For companies, publishers, and communication teams, the impact is direct. Visibility increasingly depends on how content is structured for platforms, but also on brands' ability to maintain public trust in an environment where sources rapidly mix.

This change also puts pressure on digital platforms. Recommendation systems, content moderation, and verification mechanisms are becoming critical components of the informational infrastructure, not just simple engagement functions.

Retailers demand clearer rules for AI-generated ads

Eurocommerce, an organization representing retailers such as Amazon, H&M, Inditex, and Ikea, has asked the European Union for certain AI-generated ads to be exempt from the transparency obligations outlined in the AI Act. The request targets ads created with AI that present products and are not intended to mislead the public.

The discussion stems from Article 50 of the AI Act, which introduces transparency requirements for artificially generated or modified content. Retailers argue that overly broad application of these rules could lead to excessive labeling, reducing the utility of warnings for consumers.

For the digital marketing industry, the topic is important because AI is already used for product images, virtual models, creative variations, and personalized campaigns. An unclear rule could affect production speed, costs, and how brands experiment with generative content.

At the same time, the debate shows that AI in advertising needs a differentiated framework. Not all content created with AI carries the same level of risk, but transparency remains essential where image, voice, or identity can create confusion for users.

Jeff Bezos sees AI as a driver of labor demand

Jeff Bezos stated at VivaTech 2026 that artificial intelligence could lead to a labor shortage, not a massive replacement of people. The Amazon founder presented AI as a technology that can eliminate operational barriers and enable people to do more work, not less.

The statements came in a context where many companies are discussing automation, efficiency, and restructuring associated with AI. Bezos chose a more optimistic perspective, arguing that the emergence of new technological capabilities can generate industries, products, and services that do not exist today.

For business, the message is relevant as it changes the main question. It is not just about how many roles can be automated, but about which processes can be extended, which products can be built faster, and which skills will become necessary in an AI-assisted economy.

In practice, the impact will depend on how organizations implement the technology. AI can accelerate production and analysis, but real value emerges when it is integrated into clear processes, with measurable objectives and teams prepared to work differently.

AI identifies threats, but security needs human control

AI can help security teams identify, prioritize, and analyze threats, but it should not hold the final control over security decisions. TechNewsWorld emphasizes that detection is only part of protection, and enforcement decisions must remain based on clear policies and human accountability.

AI tools can analyze large volumes of data, detect anomalies, and surface signals that would be difficult to track manually. This is useful in contexts with many alerts, complex applications, and distributed infrastructures, where response time matters.

However, the risk arises when AI-generated scores are treated as definitive decisions. In cybersecurity, a wrong decision can block a legitimate system, allow risky code to run, or affect critical operations.

For companies, the practical conclusion is clear: AI should be used as an analysis and support tool, not as an autonomous authority. Effective security requires explainable policies, verifiable controls, and people who can be held accountable for decisions made.

Commodore returns with an anti-distraction flip phone

Commodore has introduced the Callback 8020, a retro-designed flip phone built for users who want access to useful features but fewer digital distractions. The device runs Sailfish OS, promises support for most Android applications, and blocks the browser and social media apps.

The phone is positioned between a smartphone and a feature phone. It includes features such as WhatsApp, maps, hotspot, FM radio, and an easily replaceable battery, but intentionally introduces friction through the flip format and T9 typing.

For the product design area, the product is interesting because it addresses a behavioral need, not just a technical one. Users are not necessarily looking for more performance, but a more controlled way to use technology.

The Callback 8020 also shows how nostalgia can be used as a differentiation strategy in a market dominated by similar devices. The value comes not just from hardware but from the promise of a less invasive digital experience.

Microsoft ends support for Office 2021

Microsoft has confirmed that support for Office 2021 will end on October 13, 2026. The applications will continue to function, but will no longer receive security updates, bug fixes, or technical support.

The decision affects users and companies using perpetual license versions of Word, Excel, PowerPoint, or Outlook. After the announced deadline, the software may remain operational but will become increasingly exposed to vulnerabilities.

For IT departments, this type of transition must be treated as an operational risk issue. Migrating to Microsoft 365 or a newer version requires planning, compatibility audits, and checking existing integrations.

In corporate environments, productivity applications are connected with documents, internal flows, email, CRMs, and approval processes. Software without support can create risks that go beyond simple individual use.

Apple prepares the transition to M6 chips

Apple may launch the first products with M6 chips this fall, according to information published by 9to5Mac. The new chips are expected initially on Mac products, in the context where Apple continues to push and expand its strategy based on its own silicon.

The transition to M6 would continue the direction started with the M series, where performance, energy efficiency, and hardware-software integration are more tightly controlled by Apple. For users, this could mean faster and more efficient devices, especially in creative tasks or local AI.

For developers, the new generation of chips may influence how native applications are designed. Features that rely on local processing, graphics, machine learning, or multimedia workflows can benefit from more specialized hardware.

Although the information remains in the rumor zone, the direction is important for the software market. As AI increasingly moves to devices, hardware architecture differences may become a strategic criterion for developing digital products.

Vanity search enters the AI era

“In the Weights” is a new project that transforms the idea of vanity search into an AI-centered context. The platform attempts to measure how well an AI model can recognize and remember a person without using external tools like web search.

The name refers to “weights,” numerical parameters that influence the behavior of an AI model. The concept starts from the question of whether a person, creator, or professional is represented strongly enough in the data and structure of a model to be recognized without online search.

The subject is relevant for digital identity. In the past, visibility was measured by search engine results, media mentions, or social profiles. In the AI ecosystem, a new question arises: how are people and brands represented in the responses generated by models?

For companies and professionals, this type of tool shows a broader change. Digital presence is no longer just about classic SEO, but also about data, context, authority, and how public information is assimilated by AI systems.

Reliance wants AI in calls, applications, and homes

Reliance Jio has announced Jio Call Agent, an AI assistant that can join calls, transcribe conversations, generate summaries, and perform tasks such as bookings, food orders, or calling a taxi. The service is expected for over 500 million Jio users.

The initiative is part of a broader strategy through which Reliance wants to integrate AI into calls, applications, home services, and everyday digital experiences. For India, the stakes are high, as Jio has a huge user base and can transform AI into an infrastructure function, not just a separate product.

For the telecom industry, the project shows how operators can become AI platforms on a large scale. The network, applications, data, and voice interface can be combined into an ecosystem where AI actively assists user interactions.

The impact for software and digital services is significant. If such agents become part of mass calls and applications, companies will need to rethink support, booking, commerce, and integration flows with conversational interfaces.

Users remain hesitant about AI in dating

A Match survey cited by TechCrunch shows that 47% of single people in the U.S. aged 18 to 39 view the use of AI in romantic contexts negatively. However, the same research suggests that some users are open to AI features that help with profiles, photos, or conversations.

This difference is important for dating platforms. Users may accept AI as a support tool, but they do not want technology to replace the authenticity of human interactions.

For product design, the limit is clear. AI can reduce friction in areas like writing profiles or breaking the ice in a conversation, but it becomes problematic when it seems to falsify intention, personality, or connection.

The subject shows that AI adoption depends heavily on context. In utility applications, automation is often welcome. In spaces where trust and vulnerability are central, users demand more control, transparency, and authenticity.

Adobe expands AI assistant in creative applications

Adobe has expanded the AI assistant Firefly into Premiere, Illustrator, InDesign, and Frame.io. In Premiere, users can use AI for organizing materials into bins, renaming clips, identifying questions from interviews, and adding markers.

The new features are primarily aimed at automating operational steps in creative workflows. In Illustrator, the assistant can reorganize layers or check for missing fonts, while in the Firefly ecosystem, brand kits, storyboards, and product video materials can be created.

For creative professionals, AI is positioned not just as an image generator but as a tool for orchestrating tasks. This can reduce the time spent on organization, preparation, and repetitive adjustments.

For companies that produce content at scale, integration is relevant because AI enters directly into the software used daily. Productivity does not depend only on the quality of the model but on how well it is integrated into existing processes.

Global leaders want American AI, but not total dependence

At the G7 summit, leaders such as Emmanuel Macron and Narendra Modi expressed concerns about dependence on American AI models. The central theme was the risk that access to critical technologies could be rapidly interrupted by the United States.

This tension reflects an increasingly important issue in AI governance. States and companies want access to the most advanced models, but they also want guarantees regarding continuity, data sovereignty, and control over critical infrastructure.

For Europe, India, and other large markets, the subject is strategic. AI is no longer just a commercial technology but a component of economic competitiveness, digital security, and public administration.

For software providers, this context may accelerate demand for local models, sovereign cloud, hybrid infrastructure, and solutions that reduce dependence on a single provider or jurisdiction.

Anthropic joins the Frontier coalition for carbon removal

Anthropic has become the first pure AI company to join the Frontier coalition, an initiative dedicated to funding carbon removal technologies. Frontier announced a new commitment package of $915 million, raising the total to $1.8 billion.

Frontier was founded by companies such as Stripe, Google, and Shopify to support carbon removal technologies still in development. So far, the coalition has contracted nearly $700 million for over 50 projects, aiming to remove 1.8 million tons of carbon.

Anthropic's entry is notable as it comes at a time when AI companies are criticized for their increasing energy consumption and infrastructure. As models grow, data centers, energy, and cooling become central parts of the sustainability discussion.

For the AI industry, the message is that climate responsibility becomes part of the growth strategy. Companies developing advanced models will be evaluated not only by performance but also by how they manage the impact of the infrastructure they use.

Pinterest tests conversational shopping with AI

Pinterest has launched Ask Pinterest, an experimental AI-based shopping application that allows users to request recommendations and inspiration through natural language. The application extends Pinterest's visual discovery experience into a conversational interface.

Ask Pinterest can use users' saved Pins and boards to personalize responses. By separating the application from the main platform, Pinterest can test new behaviors without immediately altering the central user experience.

For digital commerce, the direction is important. Traditional keyword-based search is gradually being replaced by conversational interactions, visual recommendations, and experiences closer to a digital consultant.

For brands and retailers, this type of interface can change how products are discovered. Visual data, structured descriptions, and usage context become more important for appearing in AI-generated recommendations.

Consumers react negatively to brand messages about AI

A WordPress VIP survey cited by TechCrunch shows that 60% of consumers in the U.S. find the use of the term "AI" in brand messages unattractive. The research also shows that 86% of consumers do not fully trust AI and still want access to original sources.

The results indicate a gap between companies' enthusiasm for AI and the public's perception. For many brands, AI is presented as a signal of modernization, but for users, it can become a reason for skepticism if not clearly explained.

For marketing and communication teams, the practical conclusion is that AI should not be promoted as a value in itself. Consumers are more interested in concrete benefits, clear sources, transparency, and digital experiences that seem useful and human.

In an ecosystem where automatically generated content becomes increasingly visible, differentiation will come from credibility. Brands that use AI discreetly, responsibly, and user-oriented can gain more than those that make it a central message.

Vodafone calls for Europe to allow telecom scaling

Vodafone CEO Margherita Della Valle has urged European governments to accelerate regulatory changes that would allow telecom groups to consolidate and invest on a larger scale. The statement was made in the context of Reuters NEXT, referring to the gap with the U.S. and Asia.

Vodafone argues that European operators need larger scale to support investments in digital infrastructure. Mobile networks, fiber, cloud, edge computing, and connectivity for AI applications require significant capital and long-term planning.

For Europe, the subject is about digital competitiveness. If infrastructure does not evolve quickly enough, AI services, industrial applications, and advanced digital products may be harder to scale.

For companies, the quality of telecom infrastructure directly affects the performance of software products. Connected applications, real-time services, IoT, and AI at the edge depend on stable, fast, and well-developed networks.

Nvidia and Microsoft prepare PCs for local AI agents

Nvidia and Microsoft are collaborating to bring to Windows PCs built for local AI agents through the RTX Spark platform. Nvidia states that RTX Spark offers 1 petaflop of AI performance, up to 128 GB of unified memory, and dedicated technologies for running agents on personal devices.

The direction marks an important shift for personal computers. Until now, AI has been primarily associated with the cloud; the new generation of PCs aims to bring some processing directly to the device.

For users and companies, local processing can mean lower latency, better privacy, and more predictable costs for certain tasks. Local AI agents could manage documents, applications, files, and workflows without constantly sending data to the cloud.

For software developers, this could open a new category of applications. Digital products will be able to combine local models, cloud services, and agent interfaces in a way tailored to each user's needs.

AI expansion shifts investor attention to the bond market

Massive investments in AI infrastructure are changing how investors view large tech companies. CNBC notes that firms that recently had very strong balance sheets are now consuming reserves and using more debt for data centers and AI infrastructure.

This shift makes interest rates and the bond market more important for evaluating the tech sector. The cost of capital matters more when AI expansion depends on large expenditures for GPUs, energy, land, cooling, and building data centers.

For investors, the central question is whether the revenues generated by AI will justify the pace of investments. In the absence of clear results, enthusiasm may be pressured by higher financing costs.

For companies using AI services, this dynamic can influence pricing, availability of infrastructure, and business models. AI is not just software but also a massive effort in financial, energy, and operational infrastructure.

AI leaders send a troubling message to governments

At the G7 summit, leaders from OpenAI, Google DeepMind, and Anthropic discussed with political leaders about the potential and risks of artificial intelligence. Sky News describes the moment as tense, as companies building the most advanced AI systems warn about their risks at the same time.

The presence of these executives at the table of global leaders shows how much the importance of AI has grown on the political agenda. The technology is no longer treated merely as an economic tool but as a subject of security, competitiveness, and international governance.

For the industry, the message is ambivalent. On one hand, AI promises productivity, scientific discoveries, and new business models. On the other hand, the speed of development raises questions about control, accountability, and institutions' ability to keep pace.

For companies, this context confirms the need for internal governance. AI adoption must be accompanied by clear policies, risk assessments, data control, and mechanisms through which automated decisions can be explained and verified.

Apple limits watchOS 27 to newer models

Apple has explained that the decision to stop support for watchOS 27 for several Apple Watch models is related to processing requirements. Models such as Apple Watch Series 6, 7, 8, SE 2, and the first Apple Watch Ultra will only receive basic security updates.

The company states that the new watchOS features, including AI capabilities for Siri and a new control gesture, work best on Apple Watch Series 9 and later, Ultra 2 and later, as well as SE 3.

For users, the decision may be frustrating, especially for relatively recent devices. For Apple, however, the message is that new AI features and more complex interactions require more powerful hardware.

For the software industry, the situation shows an important reality: AI features can accelerate the fragmentation of hardware support. As applications become more dependent on local processing, older devices may exit the full update cycle more quickly.

Google DeepMind proposes a framework for controlling AI agents

Google DeepMind has published an AI Control Roadmap, a security framework for managing advanced AI agents. The approach starts from the idea that agents should be treated as potentially imperfectly aligned systems that need additional controls at the system level.

DeepMind describes the model as a defense-in-depth strategy. Alongside model alignment, sandboxing, endpoint security, resistance to prompt injection, and incremental access based on verified behavior are necessary.

For companies, this framework is relevant as AI agents begin to receive access to files, applications, databases, and internal tools. As autonomy increases, so does the risk that an agent may execute unwanted actions or be manipulated.

The message for the software industry is that AI security cannot be solved just by a better model. It requires architecture, permissions, auditing, monitoring, and clear limits on the actions agents can execute.

Estonia creates digital identities for AI agents

Estonia has announced that it will introduce digital identification codes for AI agents. The government states that the system will allow autonomous agents to act on behalf of people, companies, or organizations, with defined rights and clear responsibilities.

The initiative fits into Estonia's tradition of building advanced digital public services. Through AI ID codes, the country aims to prevent situations where an AI assistant receives too broad access to data or systems to perform simple tasks.

For business, the concept is important because AI agents will need identity, authorization, and traceability. If an agent fills out a form, submits a statement, or interacts with a public system, it must be clear who authorized it and what limits it has.

For digital infrastructure, this could become a major direction. AI agents will not be able to operate at scale in critical systems without identity, permissions, auditing, and legal accountability mechanisms.

Americans remain skeptical about the impact of AI

A Pew study cited by TechCrunch shows that only 16% of Americans believe AI will have a positive impact on society in the next 20 years. About 40% estimate a negative impact, and 67% do not believe the U.S. government will significantly regulate AI.

Public skepticism contrasts with the enthusiasm of investors and the tech industry. Even as AI usage increases, trust in companies and institutions remains limited, especially regarding safety and control.

For companies, this gap is important. AI adoption does not depend solely on technical performance but also on user perception, transparency, and how the technology is explained.

For the software industry, the conclusion is that AI must be implemented carefully. Products that ignore user anxieties risk encountering resistance, even if the underlying technology is advanced.

AI reduces the duration of technology projects at Deutsche Bank

Deutsche Bank uses AI to accelerate internal technology projects, reducing some timelines from years to a few months. Denis Roux, CIO for the investment banking division, stated that activities that used to take two years can now be completed in three to six months.

The bank uses AI to reduce backlogs, automate data extraction, support analysis, and connect market events with portfolio evaluations. India plays an important role in this strategy, with about 9,000 employees in technology roles, representing 45% of the bank's global tech workforce.

At the same time, Deutsche Bank closely monitors costs. Token-based or usage-based pricing models can become expensive, so the bank offers usage quotas to engineers and expands access only where value is demonstrated.

For large companies, the example shows that AI can produce real productivity gains, but only if managed as an operational investment. It is not enough to be introduced into teams; it must be measured, controlled, and adapted based on the type of task.

AI helps doctors diagnose rare diseases in children

Researchers from Boston Children’s Hospital, Harvard, and OpenAI have used the o3 Deep Research model to reanalyze 376 unresolved pediatric cases, identifying leads that led to 18 diagnoses. The study was published in NEJM AI.

The analyzed cases included rare and complex diseases, following previous clinical and genomic evaluations. The results included diagnoses for neurodevelopmental disorders, neuromuscular conditions, early psychosis cases, and situations associated with sudden deaths.

The impact is significant because rare diseases can remain unexplained for years, even after advanced analyses. AI can help correlate genomic data, clinical notes, and medical literature, generating hypotheses that experts later verify.

For the healthtech area, the study shows a model of responsible AI use: support for specialists, not replacement of doctors. Value emerges when the model accelerates analysis, and the final decision remains within a validated clinical flow.

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