Agregátor RSS

As Europe seeks to assert its technological independence from the US vendors Aleph Alpha, once seen as Germany’s sovereign AI hope, is the target of a transatlantic takeover.

Aleph Alpha is set to merge with Canada’s Cohere in a deal that will bring together Cohere’s global AI clout and Aleph Alpha’s background in research. The two companies hope they will be able to develop an AI powerhouse, with backing from their Canadian and German ecosystems

“Organizations globally are demanding uncompromising control over their AI stack. This transatlantic partnership unlocks the massive scale, robust infrastructure, and world-class R&D talent required to meet that demand,” said ” said Cohere CEO Aidan Gomez in a news release that artfully presents the deal as a merger of equals but that, according to a footnote, only requires the approval of the German company’s shareholders, a sure sign of a one-sided takeover.

The combined companies will be looking to offer customized AI in highly-regulated sectors including finance, defense and healthcare. By pooling their talents and offerings, theu hope to offer AI solutions to organizations according to local laws, cultural contexts, and institutional requirements.

The move comes at a time when businesses across the word are looking at non-US options as a reaction to the Trump administration’s policy on tariffs and the uncertainty caused by the war with Iran.

There have been several initiatives within Europe to counteract the US dominance. The EU’s Eurostack plan looked to make sure that major projects had a European option. Aleph Alpha was one of the companies highlighted within the scheme. The EU also launched Open  Euro LLM, an attempt to slow down the US and China’s lead in AI.

This article first appeared on CIO.

A new financially motivated hacking group tracked as BlackFile has been linked to a wave of data theft and extortion attacks against retail and hospitality organizations since February 2026. [...]

Microsoft will roll out passkey support for phishing-resistant passwordless authentication to Microsoft Entra‑protected resources from Windows devices starting late April. [...]

Microsoft has beefed up Copilot’s capabilities in Word, Excel and PowerPoint, claiming its Agent Mode will help speed up workers’ output.

The new features, announced last year, mean that Copilot can work more efficiently with Office applications, for example, understanding the richness of a pivot table in Excel or the use of animations in PowerPoint.

In tests with customers and researchers, Microsoft has learned a few things about how to improve the way Copilot is deployed, and laid out some of them in a post to a company blog. Now, it said, Copilot takes action rather than just suggesting steps — although ensuring that users maintain control. Other improvements include the ability to work with different models and better integration of Work IQ to deliver higher quality output.

Further developments are in the pipeline, Microsoft said, including improved editing for complex workflows such as finance spreadsheets and legal documents, more visibility on changes, and a more seamless integration of Copilot into the software, so that the experience for users is the same for Word, Excel, and PowerPoint.

The updates are available now and are the default experience for customers with Microsoft 365 Copilot and Microsoft 365 Premium subscriptions, the company said.

See PCWorld’s first impressions of the new Copilot agents in Word and PowerPoint.

The US Cybersecurity and Infrastructure Security Agency (CISA) does not yet have access to Anthropic’s bug-hunting AI model, Claude Mythos, even though other government agencies do, Axios reported earlier this week.

As if that weren’t a big enough slap in the face for the national cyber-defense agency, the list of those who do have access to Mythos includes several unauthorized users, according to Bloomberg News. Members of a private Discord channel specializing in seeking information about unreleased AI models, have gained access to Mythos, according to one unnamed member of the group, Bloomberg reported. “The group has been using Mythos regularly since then, though not for cybersecurity purposes,” the person told Bloomberg, supplying screenshots to back up their claim.

As a result of its fear that the powerful model could be used to identify and exploit flaws in software and online services, Anthropic has limited access to a preview of Mythos to an exclusive group of government agencies, industry groups, and software providers through an initiative it calls Project Glasswing.

Even if CISA is shut out, some government agencies do have access to Mythos, including the US Department of Commerce’s Center for AI Standards and Innovation and the US National Security Agency, which Axios said are already assessing Mythos.

A new vulnerability dubbed Pack2TheRoot could be exploited in the PackageKit daemon to allow local Linux users to install or remove system packages and gain root permissions. [...]

Tencent has updated its Hunyuan AI model, its first major release since it recruited Yao Shunyu, a leading AI scientist from OpenAI. Tencent’s Hy3 model, currently available in preview, offers improvements in areas from complex reasoning to coding.

The Chinese technology conglomerate is playing catch-up with other Chinese AI developers including ByteDance, Alibaba and DeepSeek. China is betting big on open-source AI to offer alternatives to major US players. Back in 2023, Tencent claimed its then-new Hunyuan LLM was a more powerful and intelligent option than the versions of ChatGPT and Llama available at the time.

Tencent has backed AI start-ups including Moonshot AI and StepFun, hoping that they will boost its cloud computing division. The company has also restructured its research team to improve the quality of training data. It aims to double its investment in AI to more than $5 billion this year.

Not to be outdone, DeepSeek announced its V4 Flash and V4 Pro Series, the newest versions of its LLM model. DeepSeek became an overnight hit in January 2025 with the launch of its R1 AI model and has gone on to develop other models since. It said the V4 model upgrades will offer users advances in reasoning and agentic tasks, while a new feature called Hybrid Attention Architecture improves the ability of the AI platform to remember queries across long conversations.

This article first appeared on InfoWorld.

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has revealed that an unnamed federal civilian agency's Cisco Firepower device running Adaptive Security Appliance (ASA) software was compromised in September 2025 with a new malware called FIRESTARTER. FIRESTARTER, per CISA and the U.K.'s National Cyber Security Centre (NCSC), is assessed to be a backdoor designed for remote accessRavie Lakshmananhttp://www.blogger.com/profile/[email protected]

Adobe is rolling out autonomous agents to orchestrate work across its applications, a move that will reinforce its position at the core of content and marketing workflows as AI disrupts the software landscape, analysts say.

“We’re living at true inflection point; a moment where creativity and marketing are being reshaped by AI, unlocking incredible new opportunities and raising the bar for speed, personalization, as well as scale,” said Shantanu Narayen, Adobe CEO, during his keynote presentation at Adobe Summit on Monday.

Liz Miller, vice president and principal analyst at Constellation Research, described the various agent-focused product updates at Adobe’s Summit conference this week as “an evolution of vision that brings the right AI capability into the right application.”

“The goal … is to continue evolving where and how AI is incorporated into the work of engagement,” she said.

[ More Adobe Summit 2026 coverage ]

Adobe’s recent launches indicate a “clear shift” to prioritize agentic AI investment, said Maria Bell, senior analyst at CCS Insight.

“Rather than focusing on standalone AI features, the emphasis is on building systems that can coordinate and execute work across workflows and functions,” said Bell. “Capabilities such as CX Enterprise, workflow agents and Firefly integrations point to an ambition to move from AI that supports decisions to systems that can act on them.”

Adobe kicked off its agent-related announcements ahead of the customer experience conference, unveiling its Firefly AI Assistant last week.

Using natural language prompts, the agent can autonomously carry out multi-step workflows across Adobe Creative Cloud apps such as Photoshop, Premiere, Express, and others.  Aimed at both novice and expert users, Firefly AI Assistant can also guide users through tasks spanning image, video, audio, and design. A public beta is “coming soon,” according to Adobe.

The launch of Firefly AI Assistant signals Adobe’s intent to “lead in agentic AI for creative professionals, [by] directly addressing workflow friction, usability, and the demand for multi-model flexibility,” said Keith Kirkpatrick, research director at Futurum, in a blog post last week.

“Adobe’s Firefly Assistant is a signal that agentic AI is moving from experimental pilots to production-grade tools capable of handling real creative complexity,” he said, with enterprise buyers “no longer content with simple copilots or one-off automation.”

“The ability to automate multi-step tasks and orchestrate between image and video modalities is quickly becoming table stakes for creative AI platforms,” he said.

Adobe’s main announcement during the Summit event this week was CX Enterprise Coworker; an AI agent that coordinates multi-step workflows and tasks across Adobe’s customer experience applications.

“Adobe is moving to occupy the role of an automated operating system for marketing,” said Jim Lundy, CEO of Aragon Research, in a blog post on Wednesday. “While previous AI tools acted as individual assistants for specific tasks, the CX Enterprise Coworker acts as a supervisor that connects disparate silos of information.”

Lundy said that CX Enterprise Coworker represents a “significant evolution” in the way enterprises will manage the customer lifecycle, replacing manual hand-offs with automated orchestration across customer engagement apps.

“By anchoring this tool in its robust experience platform, Adobe is making a strong case for being the primary intelligence layer in the modern marketing stack,” said Lundy.

There were also updates to GenStudio with Brand Intelligence, a data layer that connects information across Adobe tools to provide context for agents to act upon, alongside a new agent capability in Adobe’s Workfront work management app.

While agents present an opportunity for Adobe, it also faces potential disruption from both design software vendors that build AI into their products and general-purpose AI assistants.

This risk has raised concerns in financial markets, and Adobe announced a $25 billion share buyback scheme this week — a move that can be seen as an attempt to shore up investor confidence amid a period of significant change, both across the industry as AI reshapes the software landscape, and within Adobe itself, with CEO Shantanu Narayen set to step down after 18 years in charge.

Ahead of the event, popular online design platform Canva unveiled its own agentic capabilities, with users able to access various Canva tools via a conversational interface that can complete multi-step processes, such as creating “a multi-channel campaign launch.”

“Canva is focused on accessibility, using AI to simplify and automate design for a broader audience,” said CCS Insight’s Bell. “This lowers barriers to entry and puts pressure on Adobe in lighter-weight and non-professional use cases.”

And, last week, Anthropic announced Claude Design, which lets users create design prototypes and marketing assets such as “landing pages, social media assets, and campaign visuals.”

In addition, Anthropic and Canva announced an integration that brings Claude Design outputs into Canva’s app.

Miller from Constellation Research said that tools such as Claude Design are “powerful additions” to the design ecosystem, enabling non-designers to quickly prototype and test ideas using text prompts. At the same time, these should be seen as more of a starting point. Professional-level design and editing tools are still required to create enterprise-ready prototypes.

“A creator may start in OpenAI, use that output in Claude to further build out the concept, but end in Firefly to ensure enterprise safety and brand controls in a more refined, finely tuned toolset,” Miller said.

Adobe is also working with a range of AI providers to make its software available where customers prefer. This includes the ability to interact with Adobe’s Firefly creative assistant directly from Claude, for instance.

“Our strategy is to meet where the users are,” said Varun Parmar, general manager of Adobe GenStudio and Firefly for Enterprise. A user might invoke an Adobe creative agent via Claude in the morning, he said, “and then later in the afternoon decide to do deeper precision and control work that requires a professional sort of interface, which is where Adobe’s product is world class.”

“We believe that these things will coexist; depending on the use case, you’ll go in and out [of different apps],” said Parmar.

As AI model providers expand into workplace software tools, it makes most sense for Adobe to focus on its core strength of serving creative and marketing professionals, Miller said.

“The risk to Adobe is more of an ongoing challenge to stay focused on customer demand and need, and not veer off course in a never-ending horse race with models proving what can be done, as opposed to commercially safe models that deliver what must be done,” she said.

And despite some media negativity around Adobe’s ability to transition into a new era of agentic AI, Miller said, Adobe’s strategy of embedding data, assets, and workflows into the tools marketers and creatives use remains sound.

Bell sees agentic systems as a “longer-term structural shift, while the more immediate pressure comes from accessibility-focused platforms like Canva.”

Yet Adobe’s access to data and expertise serving large customers provide it with an edge. “Adobe remains strong in professional and enterprise environments, where depth, control and integration still matter,” Bell said.

Tim Cook skončí jako ředitel Applu po 15 letech. Zpočátku se expert na tabulky jevil jako zvláštní volba, protože jeho předchůdce byl inovátor. Cook ale zoptimalizoval výrobní a dodavatelské řetězce a rozvíjel portfolio, které podědil po Stevu Jobsovi. Apple dnes nevyrábí jediný model iPhonu, ale ...

Nové verze AI modelů. Společnost OpenAI představila GPT‑5.5. Společnost DeepSeek představila DeepSeek V4.

Silicon often from US, but the kit from APAC and elsewhere

America's telco regulator has clarified its ban on foreign-made routers also includes mobile hotspots and domestic routers that use a 5G cellular connection to the internet.…

Leak-site bragging meets breach hunters as Have I Been Pwned flags millions of records

Carnival Corporation, the world's largest cruise company, is dealing with choppy waters after Have I Been Pwned flagged what it claimed were 7.5 million unique email addresses all allegedly tied to one of its subsidiaries. …

Latest in long-running pwning of Cisco kit found in mystery Fed agency

A US federal agency was successfully targeted by a previously unknown backdoor malware called Firestarter, according to CISA cybersnoops and their UK counterparts – neither of which disclosed the agency's name.…

Technology promised simplicity. It delivered complexity. AI promised resolution. It is delivering acceleration. The paradox is not a bug. It is the feature. The question is what we choose to do about it. *** Every new technology arrives wearing the costume of liberation. It promises freedom. It promises options. It promises to save us time, […]

One way to deal with bug hunting LLMs: ditch the old drivers

One tactic to deal with LLM-powered vulnerability detection is simple – just speed up the removal of old code. If it's gone, it no longer matters if it's buggy.…

The Office of Inspector General (OIG) of the U.S. National Aeronautics and Space Administration (NASA) has revealed how a Chinese national posed as a U.S. researcher as part of a spear-phishing campaign to obtain sensitive information from the space agency, as well as from government entities, universities, and private companies, in violation of export control laws. "For years, NASA employees Ravie Lakshmananhttp://www.blogger.com/profile/[email protected]

Article 9 of DORA makes authentication and access control a legal obligation for EU financial entities. Here is what the regulation requires, and what a breach looks like when those controls are missing. [...]

Evropský parlament udělal zásadní krok k digitálním dokladům od vozidel. Nová pravidla mají ulevit řidičům od byrokracie a zatočit s podvody při prodeji ojetin.