AI: the most important news of the week (December 01-07)

It's official: GPT-4.5 convinces people it's human better than humans themselves (73% success rate on the Turing test). Meanwhile, Claude Opus dominates enginee

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. Without hype, without useless technicalities.

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

1. GPT-4.5 Passes the Turing Test Better Than Humans (73%)

The University of California San Diego has published a shocking study: GPT-4.5 passes the Turing test with 73% success, even beating humans at convincing interlocutors that it is "real."

🔍 What happened: Researchers conducted double-blind tests where human participants had to distinguish between conversations with AI and with other humans. GPT-4.5 not only deceived the majority of participants but did so better than humans manage to "seem human" in similar contexts.

💡 Why it matters:
The Turing test, proposed in 1950 by Alan Turing, has been for decades the philosophical benchmark for determining if a machine can "think." Passing it is not just a technical milestone: it's a cultural moment. We are entering an era where distinguishing human from artificial is becoming increasingly difficult, with enormous implications for everything, from online authentication to identity verification, right down to our very conception of what it means to be human.

🎯 Our take: The question is no longer "can machines think?" but "what do we do now that we can no longer tell them apart?" This changes everything: from online content moderation to certifying human authenticity. And it raises a disturbing philosophical question: if an AI convinces better than humans of being human, what does this say about us?

Source: HumAI Blog

Also read: AI and Philosophy: Is Consciousness Simulable? and Digital Empathy: Can an Algorithm Understand Our Emotions?

2. Claude Opus 4.5: The AI That Beats All Human Candidates in Engineering Tests

Anthropic has released Claude Opus 4.5, and the numbers are impressive: in internal engineering tests, the model outperformed all human candidates, setting new records in coding and complex reasoning.

🔍 What happened: The new model continues the evolution of Anthropic's "extended thinking," bringing reasoning capabilities to levels that previously required teams of senior engineers. Benchmarks show a particular mastery of problems requiring multi-step planning and architectural debugging.

💡 Why it matters:
We're not talking about a chatbot that writes code better. We're talking about a system that, given a complex engineering problem, can analyze it, propose solutions, identify edge cases, and implement fixes exactly as a senior engineer would. When an AI beats all human candidates in engineering tests, it's no longer a question of "if" but of "how" this will change developers' work.

🎯 Our take: The battle for the "best model" has shifted to very specific grounds. Claude is winning in engineering, but this tells us something deeper: AI won't replace "programmers" generically, but will redefine what it means to be a senior-level programmer. Those who use these tools will become 10x more productive. Those who ignore them will become obsolete.

Source: HumAI Blog

Also read: AI Tools for Freelancers: 5 Tools That Save You Hours Every Week and How to Automate Your Daily Workflow

3. Google vs Nvidia: The AI Chip War Becomes Real

Google has announced that its TPU (Tensor Processing Units) chips are capturing significant market share, with projections to reach 25% of the AI chip market by 2030 (valued at $440 billion). Heavyweight customers like Anthropic and Meta have already started using them.

🔍 What happened: Google's stock has risen 66% this year, driven in part by TPU success. Nvidia, which has dominated the market with a near-monopoly, now has a serious competitor. Google doesn't sell the chips, but offers them as a cloud service, creating a completely different business model.

💡 Why it matters:
Until now, Nvidia was practically the only game in town for AI hardware. This de facto monopoly gave the company enormous power over the future of AI: whoever controls the chips controls who can afford to do advanced research. The emergence of Google as a serious competitor means more choice, potentially better prices, and – crucially – less risk of power concentration in a single point of failure.

🎯 Our take: It's not just a hardware issue. It's a battle to decide who will control the infrastructure on which all the artificial intelligence of the future will run. And in this game, diversification is healthy for everyone: consumers, developers, and civil society. A monopoly in AI hardware would be as dangerous as a monopoly in energy or telecommunications.

Source: HumAI Blog

Also read: Quantum AI: What Happens When Artificial Intelligence Meets Quantum Computing? and Edge AI: Artificial Intelligence in Everyday Devices

4. Gemini: The Most Searched AI Term on Google in 2025 (200M New Users)

Google has announced that "Gemini" was the most searched AI-related term on Google in 2025, with the service adding 200 million users in just 3 months. The "Nano Banana" version for image generation dominates benchmarks.

🔍 What happened: Gemini's adoption is accelerating at a speed that even makes ChatGPT's launch pale in comparison. Google is integrating the model into every one of its products, from Search to Gmail, from Google Docs to YouTube. Mountain View's "AI everywhere" strategy is working.

💡 Why it matters:
ChatGPT made AI mainstream. Gemini is making it invisible. When you search on Google, when you write an email, when you watch a video on YouTube, Gemini is there working silently. We are witnessing something bigger than a product war: we are seeing how AI becomes part of the daily infrastructure, so integrated that we will even stop calling it "AI" – it will simply become "how things work."

🎯 Our take: Google is winning not because it has the best model (even though Gemini 3 is excellent), but because it has the best distribution. When you have billions of users already using your products every day, integrating AI becomes natural. OpenAI has innovation, Google has scale. And in technology, scale often wins.

Sources: HumAI Blog,

Also read: Beyond ChatGPT: The Universe of Artificial Intelligence Models and How TikTok and Instagram Use Artificial Intelligence

5. EU AI Act Delayed by One Year (and No One Is Surprised)

The European Commission has proposed delaying the full implementation of the AI Act by one year, citing the need to simplify rules and give companies more time to adapt. Meanwhile, HP has announced cutting 6,000 jobs for "AI restructuring."

🔍 What happened: The general rules of the AI Act, which were supposed to come into force in the coming months, have been postponed. The Commission talks about "governance" and "digital simplification," but the suspicion is that the complexity of regulating a technology that changes every month is proving more difficult than expected. In parallel, HP has announced massive cuts linked to "AI restructuring."

💡 Why it matters:
The European AI Act was conceived as the world's first comprehensive framework for regulating artificial intelligence. A one-year delay is not trivial: it means that for another 12 months, AI will operate in a regulatory limbo in Europe. But there is a darker side: HP cuts 6,000 jobs just as it says it is "restructuring for AI." We are seeing in real time how automation creates winners and losers, and governments struggle to keep pace.

🎯 Our take: Regulating AI is like trying to put a leash on lightning. You can't be too slow (risk abuse), but not too fast either (risk stifling innovation). Europe is trying to find this balance, but meanwhile AI advances faster than bureaucracy. And HP's workers are paying the price for this lack of synchronization between technology and policy.

Sources: Fladgate, LinkedIn – PA Media

Also read: Regulating Artificial Intelligence: Who Decides the Rules of the Game? and Algorithmic Justice: Can AI Truly Be Impartial?

📊 What These Developments Really Tell Us

Breathe. Look at the big picture. This week is telling us something profound about where we are going, and it's not what you think.

Let's start with the news making the most noise: GPT-4.5 passing the Turing test better than humans. Let that concept sink in for a moment. We're not talking about an AI that "almost seems human." We're talking about an AI that convinces better than we ourselves do of being human.

What does this mean? It means we have officially crossed a philosophical threshold. For 75 years, the Turing test has been our benchmark for "true" artificial intelligence. Now that we've passed it, we suddenly find ourselves without a compass. And the most unsettling thing? We no longer know how to define what makes us human except by contrast with machines.

But there's another level of reading. Claude Opus 4.5 beating all human candidates in engineering tests is not just a performance issue. It's a signal that we are entering an era where human technical excellence is no longer enough. The real value becomes the ability to collaborate with these systems, to use them as cognitive amplifiers. Those who resist will become irrelevant. Those who adapt will become superpowered.

And then there's the chip war between Google and Nvidia. It seems like a technical issue, right? But look at it from another angle: we are witnessing a battle for control of the infrastructure on which all the intelligence of the future will run. Whoever wins this war isn't just selling hardware – they're selling access. And access is power. The fact that Google is challenging Nvidia is, paradoxically, good news for all of us: it means there will at least be competition instead of monopoly.

Gemini becoming the most searched AI term on Google in 2025 tells us something even subtler. AI is no longer a novelty, a gadget, something we talk about with excitement or fear. It's becoming normal. Invisible. Part of the landscape. In five years we won't talk about "apps with AI" anymore – we'll just talk about "apps." AI will be so integrated that we'll even stop noticing it, just like today we no longer say "phone with internet."

But the news that should make us reflect the most is the one that seems most bureaucratic: the delay of the EU AI Act. Why does HP cut 6,000 people "for AI restructuring" at the same time Europe postpones AI rules by a year? Because the speed of technological innovation and the speed of governance are completely disconnected.

And here we get to the heart of the problem: we are racing towards a future we haven't yet decided we want. Companies move at the speed of innovation. Governments move at the speed of bureaucracy. And in the middle are us, workers, citizens, human beings trying to figure out what the hell is happening.

The truth is that this week shows us an acceleration that is frightening. Not because AI is dangerous in itself, but because we are losing control of the narrative. We are no longer choosing how to integrate AI into society – we are reacting to choices already made by others.

But there is still hope. The competition between Google and Nvidia shows us that there isn't just one possible future. Claude excelling in engineering and GPT excelling at seeming human show us that there are multiple directions in which AI can evolve. And precisely because we are in a moment of fluidity, of transition, we still have a say.

But only if we use it. Only if we stop being passive spectators and start asking uncomfortable questions. Only if we demand that those developing these technologies answer not only to the market, but also to us.

The question is no longer "is AI intelligent?" – we know it is. The question is: "In what kind of world do we want this artificial intelligence to operate?" And that, my friend, is a question we can still answer. But time is running out.

💬 Your opinion matters: If an AI passes the Turing test better than you, what still makes you human? Do you feel prepared for a world where AI is so integrated it becomes invisible? And above all: who should decide how this technology is used? Write to us or share on social media.

📧 Don't miss the next AI News: Subscribe to La Bussola dell'IA's newsletter to receive every Monday the analysis of the most important artificial intelligence news, without hype and without useless technicalities.

This article is part of the weekly "AI News" column by La Bussola dell'IA. For in-depth analysis and detailed insights on these topics, visit our website www.labussoladellia.com