Meanwhile, Europe is investing heavily and a gap is opening up in the workplace between how skilled employees appear and what they can do themselves.

Every two weeks, we publish an overview on the Xylos blog of what is moving in generative AI, with the necessary context. It is based on the LinkedIn overview by Tom Van ’t veld, Learning Innovator at OASE (powered by Xylos). Tom follows the developments, we translate them into what they mean for organisations. Welcome to edition eleven.

OpenAI released GPT-6 Astra. The model required so much computing power that usage for Pro subscribers was temporarily paused. Anthropic launched Claude Fable 5.1, which according to early tests codes better and is between 25 and 45 percent cheaper. Closer to home, the Ghent police deployed an AI chatbot that helps citizens day and night.

For organisations, that price drop may matter more than yet another benchmark win. Choose the model per task that fits the cost and quality you need. Deploying a chatbot for customers? Then decide in advance how you catch wrong answers and when an employee takes over.

Researcher Jacob Coxon, who worked at OpenAI and Anthropic, resigned. According to him, the builders fear that self-improving AI could wipe out humanity by the end of the decade. Anthropic CEO Dario Amodei also argues for a slower pace, so that models can be better controlled. OpenAI chief scientist Jakub Pachocki admits that nobody is ready for what is coming.

American politics reacted in a divided way. Senator Bernie Sanders demands a pause on advanced AI. President Trump mainly wants to keep the lead over China. A global agreement is therefore not on the horizon. As an organisation, draw your own lines, with an AI policy that defines which applications you allow and where human oversight remains mandatory.

French company Mistral raised three billion euros, the largest private funding round ever for a European tech company. Poland and the Czech Republic are jointly building AI data centres with European support. The EU is placing ChatGPT, Reddit and Roblox under the strictest DSA supervision. Nvidia, for its part, is acquiring Hugging Face, the central platform for open models.

The playing field is shifting fast. Keep your AI stack flexible, with a well-considered mix of vendors and a plan for when a player changes course.

Research shows that many employees hide their use of AI. They get praise for work that was largely produced by the software. That becomes a problem when you base decisions on expertise that actually came from an AI tool. Meanwhile, the state of New York is looking for a balance in education: no generative AI for the youngest pupils, but AI training for older students.

Hidden AI use often points to missing agreements. Make AI use open to discussion and invest in skills, so employees can assess AI output instead of simply passing it on. With Learning Services, we help organisations build those skills.

To close, something from our own soil: a Belgian network box promises to filter out phishing and AI slop locally. There is something poetic about needing a little device to protect us from the output of devices we thought were so great.

We will be back in two weeks with the next edition.

Tom has worked at Xylos for years. He started as a Microsoft Office trainer and grew into the driving force behind innovative learning concepts. He co-founded OASE, the online learning platform of Xylos, and PlayForward, the new gamified learning brand of Xylos. He also developed the Digital Coach concept, a Microsoft Teams Escape Room app and the mAIndset game, which teaches employees to prompt AI in a playful way. As Learning Innovator, he has increasingly focused on what AI means for the way we learn and work. Want to respond or continue the conversation? Find him on LinkedIn.