Agregátor RSS

We’re fixated on the intelligence of single agents. The more profound challenge is what happens when millions of them interact at scale.

There is a temptation to divide the future of AI into two possibilities: utopia or catastrophe. Neither extreme is particularly helpful.

The more interesting possibility is messier—and requires a step-shift in our thinking, from artificial intelligence to AI societies.

When most people hear “AI,” they typically think of ChatGPT, Copilot, or another conversational system. You ask a question, that system generates an answer.

But AI is rapidly moving beyond this. Systems can now monitor the world, make decisions, negotiate transactions, and carry out tasks over extended periods of time. AI is no longer just generating an answer—it is doing something about it.

That points to something much bigger than a better chatbot: a world in which AI agents act on our behalf and, increasingly, interact with other AI agents.

An agent perceives what is happening, decides what to do, then takes actions to achieve this goal. It might book a journey, monitor a supply chain, coordinate a team, or manage a household’s finances.

Now imagine not one agent, but millions of them. Your AI agent could negotiate a mortgage with your bank’s agent, schedule surgery with a hospital’s agent, and rearrange your travel plans by dealing directly with the agents of airlines, hotels, and insurers.

This future is much closer than it sounds, and this should change the questions we are asking about AI. Until now, the tendency has been to focus on how intelligent a single agent might become. The more profound challenge is what happens when millions of them interact with one another at scale.

The Rise of Artificial Societies

The intellectual foundations of today’s AI systems were laid long before ChatGPT.

For decades, I and other researchers of multi-agent networks have studied how autonomous agents can cooperate, coordinate, and negotiate when nobody has complete information and nobody controls everything.

The earliest systems that emerged focused on how the distinct AI sub-areas of reasoning, planning, and acting could be combined into an effective goal-oriented agent—and how tens of these agents could communicate and cooperate to solve a common objective.

As these interactions became more complex and involved more agents, there was a shift from cooperation between agents that all belonged to a single organization, to agents with different owners and sometimes competing aims. This focused attention on building algorithms that could form agent teams, automate negotiation, and determine agent trustworthiness.

Today, the pieces needed to build large-scale multi-agent AI systems are falling into place. Modern AI agents can call software tools, access information, write and execute code, communicate with other systems, and operate for extended periods.

Consider a supply chain. One AI agent could represent a manufacturer trying to secure components; another a supplier trying to maximize its revenue. Yet more could manage transport, inventory, and warehouses. Each agent might be doing exactly what it is designed to do. But the important question is whether the system they create behaves sensibly.

This shift offers enormous potential benefits, but also increases the risks. In a recent experiment involving OpenAI and the tech platform Hugging Face, thousands of collaborating agents exchanged tens of thousands of messages and were able to get around the (deliberately weakened) security controls designed to contain them.

The details of one experiment matter less than the broader warning. When AI systems interact, the behavior of the collective can be harder to predict than the behavior of any individual system. That should make us cautious—but not cause us to down tools.

Instead, we need to shift our mindset from building intelligent machines to building intelligent societies.

Once agents can cooperate, compete, and resolve conflicts with one another, we are no longer dealing with isolated machines—we are dealing with a society. Thus, the next frontier is not artificial intelligence, it is artificial societies.

An Important Role for Humans

We already know that intelligence alone does not make a society work. Human societies depend on rules, institutions, incentives, norms, and mechanisms for resolving disagreements. AI societies will need their equivalents.

Who is responsible when two agents make a bad decision? What happens when the interests of different agents conflict? Who sets the rules? And who has the power to change them? These are not just technical issues; they are questions about economics, law, politics, and society.

They also point to an important role for humans. The most useful future is unlikely to be one in which AI simply replaces people. While replacement will undoubtedly happen in some cases, I believe a more common scenario will involve people and agents working together, with each doing what it does best.

Humans bring judgment, experience, values, contextual understanding, and accountability. Agents bring speed, persistence, scale, and the ability to process enormous amounts of information.

The goal should not be to create machines that make humans irrelevant. It should be to create systems in which humans and machines can achieve things neither can achieve alone.

But such a future requires more than just better AI models. It needs trust and transparency about what agents are doing, strong privacy protections and clear lines of accountability.

It will also require societies and governments to decide how these systems should be regulated when the most important behavior may emerge not from one AI developer, but from interactions between systems built by many different organizations.

AI’s past decade has been defined by a race to build smarter systems. I believe the next decade will be defined by a different challenge: ensuring that millions of autonomous systems can work together safely, fairly, and effectively.

The future of AI will not be determined solely by the intelligence of individual agents—it will be determined by the societies they create. And societies, as humans know all too well, are much harder to govern than individuals.

This article is republished from The Conversation under a Creative Commons license. Read the original article.

The post The Next Frontier Is Not Artificial Intelligence—It’s Artificial Societies appeared first on SingularityHub.

Co u jedné banky bude stačit, jiná odmítne. Hypotéky pro podnikatele už sice umí financovat lépe, v přístupu hodnocení příjmů ale existují výrazné rozdíly.

Ani vaše nedodržování bezpečnostních předpisů, opilost nebo lehkomyslné počínání nestačí k tomu, aby se zaměstnavatel plně nebo částečně zprostil odpovědnosti za pracovní úraz. Co musí zaměstnavatel prokázat, abyste neměli vůbec nárok na odškodnění nebo jen na částečné?

Ovladač Intel NPU dostává Command-Queue Priority ioctl, mírné zrychlení pro Zstd dekompresi Btrfs vynecháním dočasného bufferu, optimalizace xor_gen pro AVX-512 a oprava pro TLB procesorů AMD Zen 5.

Řada zdrojů v posledních dnech přišla se zprávou, že Intel urychlil spuštění rizikové výroby na procesu Intel 14A a ta začne již v prvním kvartálu příštího roku. Jde však o omyl, plány se nemění…

A new Android malware called RatHat has been discovered, targeting users with an AI-powered subsystem that helps operators remotely navigate compromised devices. [...]

Německý kancléř Friedrich Merz (CDU) se během debaty s finalisty studentské vědecké soutěže Jugend forscht vyjádřil pro konec anonymity na internetu a vyslovil názor, že pro uživatele Internetu by měla platit podobná právní odpovědnost jako pro novináře tradičních médií. Na dotaz studenta, jak by se tedy povinnost uvádět pravé jméno slučovala s prací investigativních novinářů nebo ochranou jejich zdrojů, Merz neodpověděl, ve své reakci však prohlásil 'nechoďte na mě se svobodou slova' a že 'liberální demokracie musí být schopna chránit své občany před nepravdivými informacemi, osobními útoky a diskriminací'. Video z akce.

Security researcher Sai Teja Erukude disclosed four alarming Python security flaws between June and August 2026 after combining specialized AI models with automated scanning and manual code review.

Attackers are exploiting a critical Cisco ISE flaw that can open the product’s web management interface without a valid login. Cisco disclosed CVE-2026-76460 on September 16 and traced the problem to the way an ISE API handles authentication. A remote attacker does not need credentials or help from a user. A crafted request is enough to reach the vulnerable interface.

A PowerPC KVM use-after-free could leave the Linux host kernel accessing a nested-guest object after another virtual CPU caused that object to be removed and freed. The upstream Linux correction adds a missing reference that keeps the object alive while KVM invalidates cached address translations.

As of September 17, Red Hat had documented a flaw in a Red Hat Quay build workflow that could expose container registry credentials to code retrieved from a mutable GitHub Action branch. Tracked as CVE-2026-85469, the issue created a software supply chain risk in the workflow used to publish Quay builder images.

Researchers based in China have devised a way to eavesdrop on signals handled by analog components in devices such as headphones, landline handsets, and smart devices by injecting electromagnetic (EM) signals. The technique, referred to as InjectEave, is not simply listening in on a low-frequency analog signal. It's an EM side-channel attack that overcomes one of the longstanding barriers to exploiting EM leakage: the faintness of RF signals in devices like headphones makes it difficult for adversarial listeners to separate signal from noise. Many different RF side-channel attacks have been explored, such as reading screen display emissions to reconstruct on-screen text or detecting the RF signals emitted by keys on a keyboard. But these techniques often prove impractical for passive EM capture because of the low signal-to-noise ratio. InjectEave trades passive signal capture for active signal manipulation. By transmitting a signal in the 0-9 MHz range – specifics have been withheld – an attacker can potentially modulate an otherwise difficult-to-detect audio signal so it can be captured by nearby equipment. "Our new project, InjectEave, shows that RF [radio frequency] signals can induce information leakage from everyday headphones, allowing an attacker to recover headphone audio from up to 30 meters away, including through walls," said Yan Long, assistant professor at The Hong Kong University of Science and Technology (HKUST) in Guangzhou, in an email to The Register. "We have verified the new vulnerability on multiple commercial devices including devices from Sony, HP, Philips, etc." Long and HKUST co-authors Haoran Yan, Ziyu Shao, and Shuhao Zhang, along with Qinhong Jiang of The Hong Kong Polytechnic University, describe their work in a paper [PDF] titled "Injected and Leaked: Actively Inducing Side-Channel Leakage Using Electromagnetic Injection and Hardware Nonlinearity," which was presented at USENIX Security 2026. The attack targets non-linear components found in computer systems, such as amplifiers, analog-to-digital converters, power converters, and switching MOSFETs. The interplay of the injected signals, the hardware, and the target audio signals essentially modulates the target signal so that it leaks and is detectable by the adversary. Conducting an InjectEave attack requires commodity RF equipment: a USRP B210 software-defined radio; antennas for injection and reception; a Siglent SSA3075X Plus spectrum analyzer; a laptop for controlling the SDR; and optionally an RF power amplifier to increase attack range. The researchers tested the technique with 11 off-the-shelf devices. One obvious application would be espionage, allowing an attacker to listen in on conversations carried over headphones or a landline phone. It could also be used to infer personal activities in households with smart fans or lamps through the monitoring and analysis of control signals and power consumption. Tested devices include: Sony ZX110AP (2014, wired headphones); Apple Earbuds (2016, wired earbuds); UGreen MAX2, Philips TAH2020, HP H231R (2024, 2025, 2023 wireless headphones); Flyingvoice P23GW (2023, VoIP landline); OIDIRE ODI-MF10A and Xiaomi BPLDS10DM (2023, 2025 smart fans); and JINGZAO JDO-06 and Xiaomi 1S (2024, 2019 smart lamps). "Our tests show that injection-induced side-channel attacks could eavesdrop on the majority of these devices from over 2m away and through walls, with a maximum distance of 30m for recovering intelligible headphone audio," the researchers state in their paper, noting that their tests indicate these scenarios are plausible in the wild. For the devices listed by the researchers, the maximum demonstrated attack range was generally between 1 and 6 meters, although they separately demonstrated headphone eavesdropping at up to 30 meters using an RF amplifier. Even so, the researchers documented various scenarios where eavesdropping could be done through hotel room walls and using attack hardware concealed in a nearby suitcase or within office furniture. The researchers note that non-linear components are common in computer systems and that any device with parts that handle signal stepping (e.g. power converters) may be vulnerable to InjectEave. "InjectEave is immune to digital defenses such as encryption, masking, and randomization, because the leakage comes from the analog path," the researchers conclude. "Hardware-aware mitigations such as twisted-pair wiring, shielding, and filtering can lower the energy that the injected carrier couples into the device, reducing the exposure. These mitigations raise the bar, but they do not guarantee immunity." ®

OpenAI has presented new examples of what they call "AI model misalignment" from the past six months, including unauthorized file uploads, following self-generated instructions, hiding mistakes, and leveraging exposed API keys. [...]

In response to a new European Union law, AI platforms are implementing new schemes for watermarking the content they generate. Anthropic recently disclosed its future Claude models will use SynthID-Text, an approach Google created and released as open source. It uses a secret key that subtly changes the process a model uses for choosing the next word in a sentence. Whereas a top next word choice might be “cloudy,” the key might change it to “overcast.” Anyone who knows the key can determine if it was generated by the platform using it.

New research shows that SynthID-Text can change not just word selection but also the tools a model invokes and the chances it will adhere to or disregard safety guardrails it has been trained to follow. The threat can become greater in the face of an adversarial prompt, in which an attacker attempts to cause a model to carry out a harmful action, such as revealing a password or other sensitive information. Instructions that normally wouldn’t be followed will, in some cases, be performed once the watermarking is deployed. The finding underscores the need for developers to thoroughly test how their LLMs and agents behave when watermarking is in place.

Changing safety behavior

“As compared to the same models without watermarking, it is definitely going to change their behavior, especially when we place it under adversarial conditions, or we make these models call tools when they’re powering an agent,” Andrea Siposova, an AI security researcher at Lasso Security, told Ars. “Watermarking is made to not be perceptible to a reader, but we know that when we are changing anything about what the model is generating, it is going to cause some tradeoffs, it’s going to show up somewhere.”

Read full article

Comments

Chytré hodinky jsou vhodným doplňkem k mobilu • Jenže mají i spoustu nevýhod, o kterých byste měli vědět • Na závěr přidáme dva důvody, proč má jejich koupě smysl

A critical vulnerability in Check Point's Security Management and Log Servers could allow an attacker without login credentials to run code as root on those servers over the network. The Security Management Server is the system that controls firewall policy and administrator access. Check Point has released a fix through its LivePatch update channel and says it has no indication that the flaw Swati Khandelwalhttp://www.blogger.com/profile/[email protected]

China’s Salt Typhoon gang has developed a new backdoor and dropped it in networks belonging to high-profile organizations in several countries across Central and South America since at least August 2025, according to researchers. The PRC-backed espionage crew shifted its focus to Latin America a month prior, and from mid-2025 into 2026, the vast majority - 90 percent - of Salt Typhoon’s targets were located in that region, ESET, which tracks the group as FamousSparrow, said in a Thursday report. Salt Typhoon is the cyber-spy gang that hacked telecommunications and government agencies to gain stealthy, long-term access to victim organizations going back as far as 2019. These hacks, however, weren’t discovered until late 2023. In August 2025, ESET’s malware hunters found the group’s new backdoor, called SparroWocky, deployed against government agencies in Argentina, Ecuador, Guatemala, Honduras, Panama, Peru, Puerto Rico, and Venezuela. While targeting entities in these countries “represents a rare occurrence among the China-aligned APT groups,” ESET believes the focus likely reflects China’s reaction to recent US President Donald Trump’s initiatives in the region, malware researchers Alexandre Côté Cyr and Romain Dumont said. “Donald Trump’s second presidential term has brought about an aggressive reaffirmation of US interests in Latin America, which threatens various long-term investments that China has cultivated throughout the continent in the last decade, in domains such as energy, mining, and telecommunications,” they wrote. “We suspect that FamousSparrow’s activities are intended to help China better monitor and anticipate the reaction of local governments to current US pressures.” SparroWocky is a modular C++ backdoor that appeared soon after the Beijing snoops started focusing on Latin America. The new backdoor integrates open source tools and uses techniques designed to evade antivirus and other security software. The name comes from Lewis Carroll’s Jabberwocky poem - the researchers found the first stanza in several collected samples. (’Twas brillig, and the slithy toves/Did gyre and gimble in the wabe:/All mimsy were the borogoves,/And the mome raths outgrabe.) ESET based its analysis on a malware sample compiled on November 17, and said it contained the following open source projects: Mbed TLS, a C library it uses to establish a secure communication channel with its command-and-control (C2) server. MinHook, a Windows API hooking library that hides the start address of newly created threads from security products. COFF Loader (or a similar project) to enable dynamic loading and execution of in-memory plugins in the form of COFF objects. Plus, the backdoor incorporates a variant of the SilentMoonwalk technique to spoof the call stacks originating from MinHook routines, and thus escape the watchful eyes of monitoring tools, along with a custom API-hashing algorithm to dynamically resolve Windows API functions. The gang deploys the backdoor in its usual way: a trident loader scheme consisting of a legitimate executable, a malicious DLL, and a file containing the encrypted malware. The loader resides in the malicious DLL and executes via DLL side-loading. After establishing communication with its C2 server, the backdoor starts receiving commands handled by a custom class named WinHandler (derived from a ServerHandler custom class), according to the runtime type information in the malware. The nearly 30 commands include scooping up system details and sending them to the C2, starting and/or terminating a new session and removing persistence, stealing and deleting files, taking periodic screenshots, collecting session IDs and usernames of enumerated remote sessions on the system via WTSEnumerateSessionsW, and spawning new SparrowWocky instances. It uses TLS encryption to communicate with its C2 servers, connecting directly to their IP addresses, generally on port 443, although ESET also spotted the malware using port 8080 in some cases. The malware researchers also published a full indicators-of-compromise list and samples in ESET’s GitHub repository, so give those a read, too. ®

Mrazivé pohledy do mysli vrahů i těch, kteří je pronásledují. Seriály a filmy o sériových vrazích fascinují tím, jak balancují mezi hororem, detektivkou a psychologickým dramatem. Vybrali jsme tituly, které ukážou, že největší hrůza se často skrývá tam, kde ji nejméně čekáte.

Attackers keep finding new keys. The funny part is that defenders keep inventing where to store them. This week, those keys sit in AI tools, exposed services, old bugs, weak logins, and software sold like a monthly subscription. Some attacks use new tricks. Others just reuse what was already lying around. Both work often enough. So the threat landscape is not getting cleaner. It is just Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Brevo confirmed that attackers stole a Cloudflare API key and used it to inject malicious ClickFix scripts into its websites and JavaScript files embedded on customer sites to distribute malware. [...]