AI News – December 8-14, 2025: When AI Goes to Space (Literally)

Training an AI in orbit? Done. An airline run by algorithms? Coming soon. Meanwhile, OpenAI faces a massive lawsuit and the EU challenges Google. But the real n

Every Monday we select and analyze the 5 most significant news stories from the world of artificial intelligence. Not just a simple summary, but a critical reading of the developments that are truly changing the industry. No hype, no unnecessary technical jargon.

Why 5 news stories? Because they are enough to stay updated without being overwhelmed by information.

1. The First AI Model Trained in Space (On Nvidia GPUs)

Starcloud, a startup backed by Nvidia, has trained the first artificial intelligence model directly in space using orbiting H100 GPUs. The goal: to create orbital data centers for real-time analysis of satellite images.

🔍 What happened: A satellite equipped with Nvidia H100 GPUs successfully completed the training of an AI model in Earth orbit. Immediate applications include fire detection, shipwreck monitoring, and real-time environmental analysis without having to send terabytes of data to the ground.

💡 Why it matters:
It seems like science fiction, but it's extreme pragmatism. Sending satellite data to the ground, processing it, and sending it back requires time and enormous bandwidth. Processing directly in space means real-time analysis: a fire detected instantly, a shipwreck identified in seconds instead of hours. We're talking about AI that saves human lives not as a future potential, but as an operational reality today.

🎯 Our take: When we think of AI in space, we imagine spaceships and humanoid robots. The reality is much more concrete: data centers in orbit processing satellite data in real time. This opens incredible scenarios for environmental monitoring, disaster management, and global precision agriculture. The cloud of the future won't just be geographically distributed, it will literally be above our heads.

Source: CNBC

Also read: AI and Ocean Exploration: The Secrets of the Abyss and AI and Climate: Can Artificial Intelligence Save the Planet?

2. OpenAI in "Code Red": 20 Million ChatGPT Logs Exposed in Copyright Lawsuit

A US judge has ordered OpenAI to share over 20 million ChatGPT conversation logs as part of a copyright infringement lawsuit. In parallel, the company has declared an internal "code red" state to improve ChatGPT in the face of growing competition.

🔍 What happened: The lawsuit involves publishers and authors accusing OpenAI of training its models on copyrighted content without permission. The 20 million logs will need to be analyzed to verify if ChatGPT reproduces protected content. Internally, OpenAI has shifted priorities and resources to ChatGPT core, postponing projects on shopping, health, and advertising.

💡 Why it matters:
This is not a simple legal battle between companies. It's the moment when the legal system must decide how copyright works in the AI era. If the authors win, it could radically change the business model of all generative AI. If they lose, we establish a precedent that training on public content is "fair use." Both scenarios have enormous implications. And the internal "code red"? It means OpenAI feels the pressure: Google with Gemini, Anthropic with Claude, even new players like DeepSeek are nibbling away at market share.

🎯 Our take: OpenAI's vulnerability is not technical, it's legal and economic. They built an empire on data "borrowed" from the web, and now the bill might be coming due. The "code red" is the right response: refocus on what ChatGPT does well instead of expanding in too many directions. But the copyright issue is a ticking time bomb for the entire generative AI sector.

Sources: LinkedIn – PA Media, Dev.to Tech Roundup

Also read: Can Artificial Intelligence Violate Copyright? Three Real Cases Sparking Debate and AI and Copyright: Who Owns the Work?

3. The First "AI-Native" Airline (IBM + Riyadh Air)

IBM and Riyadh Air have announced what they call "the world's first AI-native airline," with artificial intelligence integrated into every aspect of operations: from customer experience to personnel, from maintenance to logistics.

🔍 What happened: This isn't about adding a chatbot to the website. Riyadh Air is building the entire operational infrastructure with AI as a founding principle: algorithms to optimize routes in real time, predictive aircraft maintenance, massive personalization of the passenger experience, and AI assistants for cabin and ground crew.

💡 Why it matters:
This is the first true "AI-first" application in an ultra-regulated and complex sector like aviation. Not "we add AI to an existing company," but "we build a company from scratch with AI as its DNA." If it works, it becomes the blueprint for every traditional industry: banks, hospitals, logistics, manufacturing. If it fails, it will prove that some industries need a human touch more than we think.

🎯 Our take: Riyadh Air has an advantage: it starts from zero. It doesn't have to deal with legacy systems, entrenched processes, unions resisting automation. It can build everything optimized for AI. It's a fascinating and risky experiment. In three years we'll know if "AI-native" in critical sectors like aviation is genius or hubris.

Source: IBM Newsroom

Also read: AI in the Real Estate Sector: Automatic Valuations and Smart Investments and Smart Logistics: When AI Optimizes Deliveries

4. EU Opens Antitrust Investigation into Google for Use of Content in AI Models

The European Commission has opened a formal investigation into possible anti-competitive practices by Google in its use of online content to train AI models, with a particular focus on features like AI Overviews in search.

🔍 What happened: Brussels suspects that Google is using its dominant position in search to "suck up" content from publishers and websites to train its AI models, without adequate compensation. Worse: AI Overviews could reduce traffic to original sites, harming those who created that content.

💡 Why it matters:
Europe is doing what it does best: using regulation to limit the overreach of big tech. But there's a deeper question: who owns the knowledge on the web? Google has been indexing others' content for 25 years, now it uses it to train AI that then competes with those who created it. It's an economic short-circuit. If you win the case, publishers and creators could get royalties from AI training. If you lose, you consolidate the "content vacuum cleaner" model of big tech.

🎯 Our take: AI has created a new problem: not just "you use my content without paying me," but "you use my content to create a system that replaces me." Google will defend itself by saying it's transformative "fair use." But the EU has already shown with GDPR and the Digital Markets Act that it's not afraid to challenge Silicon Valley. This case could redefine the economics of digital content.

Source: European Commission

Also read: Algorithmic Justice: Can AI Truly Be Impartial? and Electronic Voting and AI: The Future of Digital Democracy

5. The Real AI Gap Isn't Technical, It's Critical Thinking

Fortune published an explosive analysis: the so-called "AI skills gap" is not a problem of technical skills, but of critical thinking. Executives fear that employees don't know how to evaluate AI, not how to use it.

🔍 What happened: A report on Fortune 500 executives reveals that the main concern is not "they don't know how to program with AI" but "they don't know when to trust AI and when not to." 73% of executives fear that teams lack the strategic evaluation skills needed to govern AI, not just use it.

💡 Why it matters:
This completely flips the narrative. For months we've heard "we need to train people to use AI." Now we discover that the real problem is: do they know when AI is making a mistake? Can they distinguish a brilliant output from a plausible but false one? Do they know which decisions to delegate to AI and which not to? It's a much deeper and harder problem to solve with a two-day course.

🎯 Our take: We've spent two years teaching people how to make better prompts. Perhaps we should spend the next two teaching them how to think critically with AI. Not "learn to use ChatGPT," but "learn when not to use ChatGPT." This is the skill that will determine who thrives and who sinks in the AI era. And it's not taught with a YouTube tutorial.

Source: Fortune

Also read: Artificial Intelligence and Lifelong Learning: Learning at Any Age and AI and Digital Skills: What to Learn to Not Fall Behind

📊 What These Developments Really Tell Us

Stop for a moment. Look at the whole picture. This week isn't just telling us "what's new in AI." It's telling us a deeper story about where we're going as a species.

Let's start with space. An AI model trained in orbit. When I first read it, I thought "wow, science fiction!". Then I understood: no, it's pragmatic capitalism pushed to the extreme. We're not putting AI in space because it's cool. We're putting it there because it makes economic sense. Because sending data to the ground costs. Because time is money. Because a fire detected 10 minutes earlier can save lives and forests.

And this tells us something profound: AI is no longer expanding only horizontally (more applications), but also vertically (literally, towards space) and deeply (into every aspect of our lives). There is no longer a "where" that AI cannot reach.

But then we look at OpenAI and see the flip side. 20 million logs exposed, internal "code red," legal battles over copyright. The company that seemed invincible two years ago is now in defensive mode. Why? Because it built too fast on shaky legal foundations. It took content from millions of people without asking permission, betting that "fair use" would cover everything.

And now that bet is being called. The courts will say if it was right or wrong. But in the meantime, the message is clear: the speed of innovation has surpassed the speed of the law, and now the law is trying to catch up. And when the law catches up, it's usually not gentle.

Then there's Riyadh Air, the "AI-native" airline. And here we must be honest: it's brilliant and terrifying at the same time. Brilliant because finally someone is building something from scratch with AI as a principle, not an add-on. Terrifying because we are delegating critical safety systems – aviation, literally human lives at 10,000 meters altitude – to systems we still don't fully understand how they work.

The question is not "can it work?" – probably yes. The question is "what happens when inevitably something goes wrong?" Who is responsible when an algorithm makes a decision that costs human lives? The programmer? The company? The AI itself? We don't have answers yet.

And then the EU opening an investigation into Google. Some will say "European bureaucracy hindering innovation." But let's look at it differently: Europe is asking uncomfortable questions that Silicon Valley would prefer to ignore. If Google sucks up everyone's content to train AI that then replaces those who created that content, isn't this a perfect intellectual Ponzi scheme? Take for free, monetize, and cut out those who gave you the material.

It's not a question of being pro or anti big tech. It's a question of ecosystem sustainability. If you kill all content creators because AI replaces them, what will AI train on in a few years? On content generated by other AIs? It's the snake eating its own tail.

And finally, the news that should keep us awake at night: the real gap isn't technical, it's critical thinking. It's not that people don't know how to use AI. It's that they don't know how to evaluate it. They don't know when to trust and when to doubt. And this, friends, is much more dangerous.

Because if everyone uses AI but no one knows when it's wrong, what have we built? A system where millions of people make decisions based on plausible but potentially false outputs. Where the artificial confidence of AI masks real uncertainty. Where it seems we know, but in reality we are just delegating not-knowing to a machine more eloquent than us.

The uncomfortable truth is this: we are racing towards a future where AI is everywhere – in space, in airplanes, in our daily decisions – but we are not developing the wisdom to manage it at the same speed. The tools evolve exponentially. Our ability to evaluate them critically does not.

And here lies the real risk. Not that AI replaces us. But that it transforms us into people who have stopped thinking critically, who have outsourced judgment to systems they don't understand, who have abdicated intellectual responsibility in the name of efficiency.

The week of AI in space is also the week we discovered that the real problem isn't out there, among the stars. It's in here, in our heads. In our ability