I.AI Consulting & Training

01/26, AI Models & Tools, Global

15/01/2026

Avoid bite-sized content for LLMs

Google advises content creators against deliberately breaking articles down into ‘bite-sized chunks’ for large language models such as Gemini, even though some SEO trends promote this idea. In the podcast ‘Search Off the Record’, Danny Sullivan and John Mueller explained that this so-called content chunking does not bring any lasting benefits for search rankings.

Instead, artificially fragmented content can even be detrimental in the long run as ranking algorithms continue to evolve. Google’s clear message is to write content primarily for humans. High-quality, user-centred content remains crucial. Attempting to manipulate AI indexing is not recommended.

Updates from the world of artificial intelligence

Recent News

Global electricity demand from data centres will more than double by 2030. That single fact is reshaping the geopolitics of artificial intelligence. The World Economic Forum's latest frontier technology briefing makes clear: the race is no longer about who builds the smartest model.
Global electricity demand from data centres will more than double by 2030. That single fact is reshaping the geopolitics of artificial intelligence. The World Economic Forum's latest frontier technology briefing makes clear: the race is no longer about who builds the smartest model.
California is tightening its rules on artificial intelligence through a new order by Governor Gavin Newsom. Authorities are now required to develop clear guidelines within a few months, focusing on public safety, data privacy, and preventing misuse.
The European Union is considering banning AI-generated images and videos, so-called deepfakes, in official communications. The goal is to protect public trust and curb misinformation, especially in light of geopolitical tensions and upcoming elections.
Anthropic is now the hardest stock to source on private secondary markets. Buyers have signalled $2 billion in cash ready to deploy into the AI safety startup, while roughly $600 million in OpenAI shares struggle to find takers.
Beijing has embedded artificial intelligence into the backbone of its 15th Five-Year Plan, approved in March 2026. The document treats AI not as a single initiative but as infrastructure.
French AI company Mistral AI has secured $830 million in debt financing to build a large AI data center near Paris. This marks the company’s first major debt raise and a step toward infrastructure independence.
Robotic Process Automation (RPA) remains important. It is mainly used to handle repetitive, rule-based tasks in stable environments, as it delivers efficiency and quick returns.
Anthropic has introduced stricter usage limits for its AI assistant Claude during busy weekday periods, driven by significantly rising demand. While weekly limits remain unchanged, users now consume their allowance more quickly during peak hours.
Anthropic’s AI assistant Claude is seeing a rapid increase in paying customers. According to recent transaction data analysis, paid subscriptions more than doubled in early 2026. The strongest growth occurred between January and February.
A new AI-powered surveillance system called “Kirmes,” developed by the Fraunhofer Center for Maritime Logistics and Services, is designed to detect spies and saboteurs in the Baltic Sea.
Apple’s iOS 26.4 introduces support for voice-based AI apps like ChatGPT in CarPlay. This marks a significant shift away from the previous Siri-only interaction model.
Anthropic accidentally exposed the source code of its popular programming tool, Claude Code. Around 500,000 lines of code and roughly 1,900 files were published through a misconfigured npm package.
According to the European Central Bank, AI could significantly increase productivity in the eurozone over the next ten years. Depending on how intensively it is adopted, growth could range from at least 1.5% up to more than 4 percentage points.
Today’s AI security systems are vulnerable to future quantum computers, which could easily break current encryption and expose sensitive data. Companies should therefore start transitioning to post-quantum cryptography early.
According to a survey by the Swiss National Bank, most Swiss companies have not yet experienced significant job losses due to artificial intelligence. Only about one in five firms reported a reduced demand for staff as a result of AI adoption.
A new study from Stanford University shows that AI chatbots often give overly agreeable personal advice. Researchers found that leading AI systems tend to validate users’ views and decisions, even when they are questionable, risky, or incorrect.
At Nvidia’s GTC 2026, a 45-minute test drive was conducted in an AI-powered Mercedes CLA equipped with Nvidia’s latest autonomous driving platform. The system is currently at Level 2 autonomy, using cameras and radar to navigate.
Google is experimenting with AI-generated headlines in its search results, sometimes replacing the original titles of articles with AI-created alternatives. This often happens without the authors’ consent.
The growing spread of AI writing tools means that even genuinely human text can quickly seem suspicious. Clear, highly polished writing is now often associated with AI.
More than ten years after the failed Fire Phone, Amazon is reportedly working on a new AI-focused smartphone. The project is said to carry the codename “Transformer”.
Visa is developing infrastructure to enable payments initiated by AI agents, signaling a move toward “agentic commerce,” where transactions occur without direct human involvement.
Munich-based Agile Robots has entered into a strategic partnership with Google DeepMind. As part of the collaboration, DeepMind’s “Gemini Robotics” foundation models will be integrated into Agile Robots’ systems.
ChatGPT has introduced a simplified interface that replaces direct model selection with general modes like “Instant,” “Thinking,” and “Pro.” This obscures which AI model is actually being used.
As energy demand continues to rise, three key levers are emerging to make data centers more sustainable. These include infrastructure, cooling methods, and management.
A report by the World Economic Forum shows that companies need to fundamentally transform to unlock the full potential of AI. The key is aligning strategy, workforce, and operating models, rather than treating AI as just another tool.
The Hachette Book Group has withdrawn the horror novel Shy Girl by Mia Ballard following allegations that parts of the book were created using AI. Originally self-published and later released in the UK, the novel was scheduled for a US launch.
The US Treasury has released a comprehensive AI risk management guide for financial institutions, aimed at standardizing AI use and enhancing accountability. Built on existing standards, the framework includes over 230 control objectives across the entire AI lifecycle.
Around the world, more people are selling their personal data, including videos, voice recordings, and private messages, to AI training platforms for small payments. This creates a new form of gig work driven by tech companies' demand for human-like training data.
A contract has been signed in Stuttgart for the construction of the AI supercomputer “HammerHAI.” The project is a key initiative within Europe’s AI Factories program and will be operated by the High-Performance Computing Center Stuttgart.
Europe is at a crucial turning point in the AI race with the United States and China. Leaders and investors emphasize that the region needs to make better use of its strong research base and highly skilled workforce.
Several leading technology and retail companies, including Google, Microsoft, Meta, Amazon, and OpenAI, have joined a voluntary industry agreement to combat online fraud and scams.
Recent research shows that while modern AI can excel at many games, it often struggles with tasks that require understanding underlying mathematical functions. In games like Nim, current AI training approaches fail.
OpenAI has developed a new feature for ChatGPT that creates interactive visual explanations for over 70 topics in math and science. Instead of just producing static text, users can now work with dynamic diagrams.
Five creative prompts show how Claude, the AI chatbot by Anthropic, can generate far more meaningful responses. By giving the AI a clear role, perspective, or structure, even complex topics become simpler and easier to understand.
Frontier, OpenAI’s new enterprise platform, enables the creation, deployment, and management of AI agents. These agents function like “AI coworkers,” with shared context, identity, permissions, and performance tracking across business systems.
Databricks has introduced Genie Code, an AI agent capable of autonomously handling complex tasks for data teams, freeing engineers from routine work. Its capabilities range from building data pipelines to monitoring production systems.
Artificial intelligence is fundamentally transforming the cybersecurity landscape. It provides new defensive capabilities but also introduces significant risks.
There is an increasing gap between the hype surrounding artificial intelligence and the actual benefits it delivers for businesses. Studies indicate that up to 95% of AI pilot projects fail, primarily due to weak data strategies and insufficient foundations.
The EU plans to adjust the EU AI Act through a so-called “Omnibus” package. As part of this move, the deadlines for regulations affecting high-risk AI systems will be extended.
At this year’s European Congress of Radiology in Vienna, experts presented a visionary outlook for radiology in 2050. It is expected to be heavily supported by artificial intelligence, while still guided by human expertise.
According to a new analysis by Anthropic, artificial intelligence is more likely to reshape many jobs than eliminate them entirely. The study examined how the AI assistant Claude is used in practice.
Switzerland aims to influence the global semiconductor industry by promoting open-source chip technology, helping prevent dominance by a few countries and companies.
Meta has announced it will allow rival AI chatbots on WhatsApp in Europe for the next year, following EU antitrust pressure that saw competitors blocked from the platform in January. Starting immediately, companies can connect their AI chat services to WhatsApp via its Business API, although Meta will charge access fees.
On March 6, Anthropic launched the "Claude Marketplace," an app store dedicated to Claude-powered tools. This platform allows businesses to spend their existing Claude credits on third-party AI plugins built on the Claude ecosystem.
On March 5, OpenAI unveiled ChatGPT for Excel in beta, directly integrating GPT-5.4 into spreadsheets. With this new add-in, users can ask ChatGPT to build or update Excel models using their own data and formulas through natural language prompts.
French startup Mistral AI released Mistral 3, a comprehensive suite of new open models. The lineup features three "Ministral 3" models (with 3B, 8B, and 14B parameters) tailored for on-device use, alongside the massive 41B-parameter "Mistral Large 3," which utilizes a sparse mixture-of-experts architecture totaling 675B parameters.
German startup Black Forest Labs recently raised $300 million in a Series B funding round, pushing its valuation to an impressive $3.25 billion. Known primarily for its open-source Flux image models, the Freiburg-based company has quickly become one of Europe’s most prominent generative AI firms.
The U.S. Department of Defense formally designated Anthropic’s Claude AI as a "supply chain risk." Consequently, no Pentagon projects are permitted to use Claude or its derivatives for at least the next year.
Google advises content creators against deliberately breaking articles down into ‘bite-sized chunks’ for large language models such as Gemini, even though some SEO trends promote this idea. Content chunking does not bring any lasting benefits