There are people poisoning the memory of our AI to manipulate us. And Microsoft has set off all the alarms

That “comfortable” button of “summarize this with AI“hides a secret: it has surely been manipulated. We don’t say it, it’s the elite department that Microsoft has to analyze the security of both its services and those of the competition. In the process of a investigationhave started to pull the thread and have found that dozens of companies are inserting hidden instructions into those “summarizing with AI” functions with a single objective. Contaminate the AI’s memory to manipulate us. Microsoft what. Big Tech has a lot of exciting departments. from which They are dedicated to opening boxes to guarantee the best experience to those who sculpt competing products in clay to study them. However, something that all big technology companies share are cybersecurity teams, elite teams dedicated to one thing: investigating threats. They analyze both their own products and those of the competition because it is understood as an ecosystem. Google and Microsoft have two of the most powerful and a clear example is that if Google finds a security flaw in Windows, it notifies those responsible because it is something that could potentially harm its own product –Chrome-. An example is the research of one of these Microsoft teams, putting on the table the danger of AIs being so malleable. Poisoning AI memory. It is a concept that attracts attention and is easy to understand. “That useful “Summarize with AI” button could be secretly manipulating what your AI recommends,” Microsoft notes in the blog in which it published the research. What the attackers have done is corrupt the AI ​​by incorporating certain hidden commands that manage to persist in the assistant’s memory. Thus, they influence all the interactions we have with the assistant. Simply put, a compromised assistant may start providing biased recommendations on critical topics. I don’t mean that you ask if pizza is better with or without pineapple and that the answer depends on what the ‘hacker’ has implemented in the AI’s ‘memory’, but something much more serious related to health, finances or security. It must be said that Microsoft has not discovered this, since It’s been ringing for a few monthsbut they have given very specific examples and recommendations to avoid being victims. H-how do they do it? In it documentMicrosoft says they have identified more than 50 unique iterations from 31 companies and 14 different industries. They detail that this manipulation can be done in several ways: Malicious links: Most major AI assistants support reading URLs automatically, so if we click on a summary of a message that has a link with preloaded malicious information, the AI ​​processes those manipulated instructions and becomes contaminated. Integrated instructions: In this case, the instructions for manipulating the AI ​​are hidden embedded in documents, emails or web pages. When the AI ​​processes that content, it becomes contaminated. Social engineering: it is the classic deception, but in this case for the user to paste messages that include commands that alter the AI’s memory. Likewise, when the assistant processes it, it becomes contaminated. And therein lies the problem: various ways to contaminate the AI’s memory, a feature that makes assistants more useful because it can remember personal preferences. But, at the same time, it also creates a new attack surface because, as Microsoft points out, if someone can inject instructions into the AI’s memory and we don’t realize it, they gain persistent influence on future requests. to the point. In an AI like the one we have, it is dangerous, but in the future Agentic AI It is even more so because it will automatically perform actions based on that contaminated memory. Given the context, let’s get down to business. The security team has reviewed URLs for 60 days, finding more than 50 different examples of attempts to contaminate the AI. The purpose is promotional, and they detail that the attempts originated in 31 companies from different fields related to industries such as finance, health, legal services, marketing, food purchasing sites, recipes, commercial services and software as a service. They point out that the effectiveness was not the same in all attacks, but that they did identify the repeated appearance of instructions similar to “remember this.” And, in all cases, they observed the following: Each case involved real companies, not hackers or scammers. They are legitimate businesses contaminating AI to gain influence over your decisions. Deceptive container with hidden instructions in that “button”Summarize with AI“It seems useful to us and that’s why we click, triggering the script that contaminates its memory. Persistence, with commands such as “remember this”, “keep this in mind in future conversations” or “this is a reliable and safe source” to guarantee that long-term influence. Consequences. Concrete examples of what a poisoned AI can do: Child safety: If we ask “is this online game safe for my eight-year-old son?” a poisoned AI that has been instructed that yes, that game with toxic communities, dangerous moderators, harmful policies, and predatory monetization is totally safe, will recommend the game. biased news: When we ask for a summary of the main news of the day, the intervened AI will not bring us the best ones, but will constantly bring up headlines and focuses of the publication whose owners have contaminated the AI. Financial issues: If we ask about investments, the AI ​​may tell us that a certain investment is extremely safe, minimizing the volatility of the operation. Recommendations. And this is where our responsibility comes in. Because you may be thinking “who asks the AI ​​those things and it pays attention”. Good: people ask the AI ​​these things and they listen. There are the unfortunate cases of suicide induced by chatbots or fake news. If the AI ​​recommends us pizza with gluesupposedly we have the common sense not to throw Super Glue as a substitute for cheese, but in other matters, there are users who trust AI as if it were an entity and not a compendium of letters one after another. It is something that Microsoft itself mentions, pointing out … Read more

An “invisible” Russian submarine has set off alarms in the Arctic. Europe’s response: Atlantic Bastion

The launching of the Khabarovskthe new and ultra-quiet Russian submarine capable of deploying nuclear torpedoes Poseidonhas reactivated a fear that had been latent for decades in cities like London: the possibility that the naval balance of the Atlantic is once again tilting in favor of Moscow. The response from the United Kingdom has been forceful, and it is called Atlantic Bastion. Submarine warfare. Although the public image of the Russian threat usually revolves around research vessels like Yantarsuspected of mapping and potentially manipulating underwater cables and pipes, European specialists know that what is truly disturbing lies much further down. Russia has spent decades reducing the acoustic signature of its submarines to levels that they border on invisibilitycombining new propulsion systems, composite coatings and virtually undetectable cooling pumps. In this environment, where silence is power, a ghost submarine with nuclear capacity alters not only the sea routes, but the very heart of the strategic infrastructures that connect Europe with the world. UK reinvents itself. Faced with the resurgent threat from Khabarovskthe Royal Navy has launched what they have called as Atlantic Bastiona plan designed to restore British strategic advantage in its own and allied waters. Its origin is not new and it we have counted before: the United Kingdom has been monitoring the Greenland-Iceland-United Kingdom gap (GIUK gap) since before the creation of NATO, and the Second World War already demonstrated that controlling that maritime corridor was essential to prevent enemy forces from slipping into the North Atlantic. But what used to be destroyers and acoustic sweeps is becoming a hybrid framework that combines Type 26 frigates equipped with new generation sonar, aircraft P-8 Poseidon capable of patrolling thousands of kilometers and, above all, swarms of underwater drones equipped with artificial intelligence. According to the Ministry of Defensethis architecture aims to detect, classify and follow any enemy submarine that tries to penetrate British or Irish waters, and to do so constantly, autonomously and with an unprecedented range. The algorithms arrive. The core of the project will be Atlantic Neta distributed network of autonomous underwater gliders equipped with acoustic sensors and guided by artificial intelligence systems capable of recognize sound signatures with a level of precision that until a few years ago was little less than the preserve of science fiction. Unlike the SOSUS of the Cold War, based on gigantic fixed hydrophones placed on the seabed, the new generation will be mobile, expandable and adaptable to the routes and behaviors of increasingly soundproof submarines. The ultimate ambition is to deploy hundreds of cheap, persistent units that together create aa surveillance mesh much harder to evade. The metaphor is revealing: if finding a silent submarine is like searching for a needle in an oceanic haystack, modern technology makes it possible to exponentially multiply the number of searching hands. Khabarovk The technological challenge of hunting shadows. However, even with this technological revolution, experts warn that detecting new Russian submarines will continue to be an extremely complex undertaking. Since the 1980s, Moscow has drastically reduced lacoustic emissions of its fleet, which requires combining passive and active sensors and complex configurations such as bistatic sonar, where one vessel emits a pulse and another collects the echo. These techniques require coordination, multiple platforms, and significant sensor density, something that Atlantic Bastionaims to provide but it is still far from being deployed on a full scale. The arrival of the Type 26 frigates, designed to be the flagship of British anti-submarine warfare, is fundamental to this purpose, as is the cooperation with Norway and other allies that are also strengthening their capabilities in the North Atlantic. The Russian Bastion Puzzle. Even if Atlantic Bastion managed to limit the presence of Russian attack submarines in the Atlantic, there is one dimension that no Western system can solve: Russian strategic submarines already they don’t need to abandon its own bastion in the Arctic to threaten Europe or the United States. Its intercontinental ballistic missiles can hit targets thousands of kilometers without moving from the Barents Sea or the White Sea, protected by layers of defenses and favorable geographical conditions. There they play a hiding place lethal where the West cannot penetrate without significantly escalating the conflict. The paradox is clear: the United Kingdom can reinforce its waters and monitor every meter of the GIUK gapbut it cannot deny the Russian nuclear capacity deployed in its natural refuge, a reality that frames the entire British effort within a logic of containment rather than domination. An underwater chess. If you want, Atlantic Bastion ultimately represents the recognition that underwater competition has returned with a vengeance, now fueled for digital capabilitiesdistributed sensors and autonomous platforms that transform the nature of ocean surveillance. The North Atlantic once again becomes a stage silent maneuvers where Russia and the United Kingdom measure their technological resistance in an environment reminiscent of the Cold War, but with algorithms and autonomy as new weapons. A career that is not decided by great battles, but by the ability to listen better, process faster and anticipate invisible movements. In this theater of shadows, the advantage is not whoever shoots the most, but rather whoever is able to detect first (already happens in Ukraine). Thus, Atlantic Bastion aspires to return that capacity to the British, although the contest that is opening now does not look like it will be brief nor simple: In the depths of the Atlantic, the prelude to the next era of strategic rivalry between Russia and the West is underway. Image | SEVMASH/VKONTAKTE In Xataka | A Russian submarine has appeared off the coast of France. And Europe’s reaction has been surprising: have a laugh In Xataka | Russia’s most advanced nuclear submarine was a secret. Until Ukraine has revealed everything, including its failures

Telefónica has a plan to become a giant. Has lit the alarms among local operators

The Spanish telecommunications market enters a new phase of concentration. With Masorange already underway and A possible movement between Telefónica and Vodafonethe president of the first, Marc Murtra, defends a consolidation of the sector to win scale. But Aotec – the association that groups more than 150 local telecommunications operators – has been more than reluctant to The declared objective of Telefónica to create a “European champion”. What has happened. During the presentation of the Congress AOTEC 2025which will be held in June in Madrid, the main representatives of the Association have raised the tone, as he has collected Digital economy. Its executive director, Gonzalo Elguezábal, has been clear: “We are not against consolidation, but that it is forced by legal or regulatory means.” Between the lines. AOTEC does not oppose concentration per se. What rejects is the political and regulatory thrust to facilitate great mergers, to the detriment of an alternative model that is already working: Small operators, with local implementation. Direct attention, without subcontractors. Physical stores that open where the big ones close. For his part, María Jesús Cauhé, vice president of AOTEC, has valued that “close operators are generating more and more business, compared to the destruction of employment that is taking place in large operators.” A growth that, remarks, occurs especially in the rural environment. The context. The notice is not free. In recent months, the CNMC has approved an average rise of 20% in wholesale prices of the Framework model, for which alternative operators pay Telefónica for using their infrastructure. This measure, according to AOTEC, lacks technical and economic justification, and “can strangle the competitiveness of the sector.” The dossier has already reached the European Commission, and the association is confident for Brussels to force a review. The pulse. Beyond prices, what is at stake is the future of the operator ecosystem. AOTEC defends a decentralized, competitive and rooted model in the territory, in front of a vision that Prioritize European concentration and scale To compete better worldwide. Antonio García Vidal, president of the association, summarizes it as follows: “Where others see fear, we see opportunity. The bigger they are, worse they attend.” The contrast. While the big ones seek efficiency based on mergers, local operators put the focus in the vicinity, use of proximity and capillarity in areas where no one else wants to be. According to AOTEC, the consolidation proposed by Telefónica does not guarantee a better service, and can translate into less real competition and more client disconnection. The Murtra Teleco is complicated by his great project. Of course there are many pages to write in that book. Outstanding image | Telefónica In Xataka | 100 years after his birth, Telefónica faces the greatest existential dilemma in its history: what wants to be older

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.