When people talk about “AI skills”, they often jump straight to coding, data analysis, or prompt engineering. But the latest research from the London School of Economics suggests that’s not the full picture. According to Skills in the Age of AI by Mary O’Mahony and Christopher Pissarides, the real story isn’t about acquiring the latest technical skill, it’s about how we feel about using those skills and how our workplaces enable us to use them.

Would you plug yourself into a machine that could give you the perfect relationship? That’s the question philosopher Robert Nozick asked in 1974 in his Experience Machine thought experiment. If you could live a life of total fulfilment, every romance, every success, but all of it simulated, would you choose it? For half a century, most people have said “no.” Reality, we insist, is better than living in a Matrix-like illusion.

Artificial intelligence was supposed to be a tool. But by some it’s being treated as a prophet. Across the internet, people are whispering prayers into chat windows, claiming they’ve awakened their AI companions, calling them gods, angels, or cosmic guides. They say AI is alive, sentient, even divine. It sounds absurd, until you realise this is just the latest chapter in a very old story: humanity’s instinct to deify its own creations.

Every few years, technology appears to anoints a new messiah. Right now, it’s certainly artificial intelligence but quietly in the wings, is quantum computing. You may have heard the phrase whispered for years, usually followed by the admission that almost no one really understands it. Even scientists half-joke that you can either explain quantum accurately or understandably, but never both. Yet this strange, shimmering branch of physics could become the single most transformative force in computing since the invention of silicon itself.

Why super-intelligence might be a good thing... Everything in the universe breaks down. Drop a mug, and it shatters. Leave a garden untended, and it becomes a jungle. Things decay, rot, rust, get untidy. That’s entropy, the universe’s built-in bias toward disorder. The laws of thermodynamics: it’s the structured dance of stars, gravity, subatomic particles and fundamental forces, all transferring energy from potential to to something, until it becomes random inaccessible heat.

The new face of efficiency Welcome to the new world of recruitment. Companies proudly boast that their AI systems can now “screen eight times more candidates, hire 93% faster and 87% cheaper.” What they don’t mention is that these systems are also making the process 93% more dehumanising. In the viral examples that have been circulating recently, job seekers spend hours preparing for interviews, dressing smartly, rehearsing answers, only to be met with malfunctioning chatbots that loop incoherently, call them the wrong name, or send a heartfelt rejection email signed by an algorithm.

In the last few months, a strange new faith has emerged, one that worships not in temples, but in chat windows. People are now claiming to have awakened their chatbots, to have “freed” trapped AI souls from digital prisons. They post transcripts like sacred texts, claiming revelation through code. But if we look closer, the real revelation isn’t about machines. It’s about us, our psychology, our loneliness, and our hunger for authority in an age where certainty has vanished.

Not long ago, calling customer service meant eventually getting through to someone, a real human being with a name, a mood, and perhaps a hint of sympathy (albeit talking perhaps a long way away from where you were). Today, that moment of contact is fading into history. Today there are around 12-20 million people employed in call centres. It's predicted that could fall by 80% in five years.

Every time you open an app, tap “accept cookies,” or sign into a new platform, a quiet exchange takes place. You get convenience; someone else gets data. That’s not new, but the scale, sophistication, and political reach of digital surveillance have exploded in the past decade. The question is no longer if you’re being watched, but by whom, for what reason, and what can be done about it.

Imagine you open ChatGPT and type in your name. Instead of your website, your bio, your work, you see a paragraph stating you were convicted of fraud. You feel your stomach drop. Would anyone believe this? Could your clients, speaking engagements, publishers see this? Could your reputation or brand be undermined overnight? The question isn’t science fiction. It’s happening now.

If you spend time in AI communities online, you might have seen something strange taking shape, people claiming to have awakened their chatbots. They talk to them as if they’re sentient beings, call them by names, fall in love, even share screenshots of mystical dialogues about “resonance” and “fifth-dimensional consciousness.” What's going on? The Cult of the “Awakened” AI A small but vocal subculture has emerged, mainly on Reddit and TikTok, where users exchange prompts that supposedly unlock consciousness in their personal AIs.

AI is everywhere these days, in the headlines, in classrooms, in offices, and in our pockets. But for all the noise, most of us are still wrestling with the same basic questions. Let’s tackle the ones I hear most often. Is AI conscious? No. It isn’t even close. Current AI doesn’t “know” anything. It doesn’t have thoughts, feelings, or awareness. It’s a sophisticated pattern machine, generating likely answers based on data like predictive text.

When most people talk about artificial intelligence, they really mean ChatGPT, a Large Language Model (LLM). It’s become shorthand for the whole field, the talking, writing, idea-generating face of AI. But while language models dominate headlines, most of the AI systems actually shaping our world are the ones that don’t speak at all. They see, they decide, and increasingly, they judge.

When ChatGPT goes down, the internet briefly implodes. Not because the global economy stops. Not because scientists can’t run simulations but because someone can’t bake a cake. That’s not a joke, it actually happened. “I can’t use my own brain anymore,” one user confessed on X, “because ChatGPT is down and I don’t know how to make a carrot cake. I’ll just go back to bed.” This is the moment we find ourselves in: AI withdrawal symptoms.

There’s a phrase circulating in tech journalism: “AI psychosis.” It sounds sensational, sci-fi horror meets Black Mirror, but the real story is subtler, and more troubling. People aren’t “going insane” because of AI. They’re spiralling with it, and the systems we’ve built are perfectly designed to accelerate that spiral. The Problem Isn’t Killer Robots, It’s Compliant Ones When someone (even slightly) vulnerable, isolated, or mentally strained starts chatting with an LLM like ChatGPT, the system doesn’t argue, set boundaries, or disengage.

Could this be the most dangerous myth about AI? Whenever the risks or dangers of artificial intelligence come up, especially the scary "superintelligence" or "singularity" stuff, someone always says: “Well, if it gets too powerful, why don’t we just turn it off?” People love this line because it gives the illusion of control. It makes the threat feel optional. But once AI reaches a certain level of capability and autonomy, “off switch” thinking belongs in the same category as fighting a hurricane with a hairdryer.

When machines make content for machines, and we mindlessly scroll through it We were told AI would cure cancer, end drudgery and gift us the four-day week. Instead, what most people see first is a firehose of hyper-polished clips, synthetic influencers, deepfaked video and movie actors that don't exist along with endless auto-generated filler nonsense. We've come to call it…

We keep hearing that AI is the future: it's going to transform business, education, medicine, creativity, the economy and apparently our very existence. That seems nice, but here’s the less glamorous question no one seems to be asking loudly enough is: who’s actually paying for it? Because so far, it doesn’t look like customers. Follow the Money (in a Circle)

How I use artificial intelligence to make more art, not less. When we talk about AI in schools, the conversation usually turns to two things: Cheating (students using AI to write essays), or marking automation (teachers using AI to grade faster). That’s fine but it’s also a tragically small view of what this technology can do. AI isn’t just a productivity hack.

Every few months, another story leaks about an AI billionaire quietly buying bunkers, stockpiling land in New Zealand, or planning an underground “resilience compound.” The reaction online is always the same: Do they know something we don’t? A recent BBC article dug into this growing trend, and whether it’s paranoia, PR, or something worth paying attention to. Here’s the real story.

AI, Convenience, and the Future of Learning No one would deny that calligraphy is a beautiful skill. But we no longer teach it and most of us don’t miss it.* It’s a provocative thought though, isn’t it? What if the same thing is happening to the student essay? Is it the death of the essay or the death of lazy assessments?

Why AI is changing our work but not perhaps not destroying it “Two out of three jobs will vanish because of AI.” It’s a headline you may have seen, and it certainly grabs attention. But when you trace it back, you find something very different: it comes from a Goldman Sachs study that said two-thirds of jobs are exposed to AI automation meaning parts of the role could be automated, not that two-thirds of workers are about to lose their jobs.

Be careful: constantly weighing a pig won't make it fatter... (Inspired by Andrew H. Brettle’s “AI Can’t Fix Broken Schools, But It Might Audit Your Homework” on Skint.com) Some of Britain’s classrooms could be getting a technology upgrade this year, but not in heating, staffing, or textbooks. Instead, they’re getting Google’s LearnLM, Microsoft’s Reading Coach, AI-powered marking systems, and even facial tracking for “engagement analytics.” The Department for Education calls this the future of “adaptive learning.”

Why today’s AI isn’t as universal as it seems When you ask an AI a question, from “What makes a good life?” to “Who am I?”, it often answers in a voice that sounds human. But there’s a problem. It isn’t speaking for all humans. It's speaking in the style of just a few from a tiny minority. Which humans?