The first time I heard the word ChatGPT, I was sitting in a workshop, half‑listening as someone described this “brilliant new tool.” I couldn’t tell you which version it was, only that the room felt split; half curious and half cautious with healthy dose of sceptism.
Three years later, that uncertainty has evaporated. AI is no longer a novelty whispered about in tech circles. It’s threaded through our phones, our workplaces, our public services and more recently even military operations. It is now infrastructure but there are still many organisations still grappling with whether to integrate this generation of AI into their business models. AI is now the headline at conferences, leadership boards and workplaces.
In the middle of all this, one question keeps tugging at me:
How much will this cost us all in the long term? Not just financially but socially, environmentally and ethically.
This is the story of why that question matters.
When AI Moved Out of the Shadows
For more than a decade, AI quietly powered the background of modern life; search engines, fraud detection, logistics, medical imaging. Most of us never noticed.
2011–2015: The early breakthroughs
- In 2011, IBM’s Watson beat human champions on Jeopardy! A major public moment for AI.
- Around the same time, deep learning started taking off in research labs.
- Siri which was launched in 2011, powered by AI systems (although not generative AI)
- In 2012, the “AlexNet” breakthrough in image recognition kicked off the deep‑learning revolution.
- By 2014–2015, companies like Google, Facebook, and Microsoft were using AI for translation, search ranking, speech recognition and recommendations.
AI had became infrastructure, running systems behind the scenes.
2016–2020: AI becomes embedded
- AI began powering fraud detection, logistics, energy grids and medical imaging.
- Recommendation engines (Netflix, YouTube, Amazon) became heavily AI‑driven.
- Voice assistants (Siri, Alexa, Google Assistant) normalised AI in everyday life.
Then, almost overnight, it stepped into the spotlight.
2020–2024: The generative AI explosion
This is the period most people mean when they say “AI has boomed.”
- GPT‑3 (2020)
- Stable Diffusion, Midjourney, DALL·E (2022)
- ChatGPT (late 2022)
- GPT‑4, Claude, Gemini (2023–2024)
Generative AI brought AI into the public imagination in a way earlier systems never did.
It was talking to us, drawing for us, writing for us. It became visible, conversational, creative and impossible to ignore.
https://www.sciencenewstoday.org/how-artificial-intelligence-is-transforming-everyday-life
But the more accessible it became, the more invisible its true cost grew.
The Hidden Machinery Behind “Magic”
We talk about AI as if it floats in the cloud; weightless, effortless and endlessly available. But behind every prompt sits a physical world straining to keep up.
Modern AI depends on massive computing power, energy, water and physical infrastructure.
Training a single large model can consume as much electricity as several hundred households use in a year. Data centres require vast amounts of water to stay cool, sometimes hundreds of thousands of litres for one training run.
Though AI may be digital, its impact is anything but.
Hydrological drought is already a serious worldwide issue and research shows it is likely to intensify throughout the 21st century as climate change accelerates. World enters era of ‘global water bankruptcy’ | UN News
And the UK is not immune. 2025 drought: how it developed in England – GOV.UK
Even here, in a country famous for wet weather we’ve faced repeated droughts, from 1995 to the scorching summer of 2022. UK river flows are projected to decline consistently out to 2080. – N535460JA.pdf– S. Parry et al.: Divergent future drought projections in UK river flows
Yet the UK is in a data‑centre boom. Capacity is expected to double by 2028–2030. Planning rules have been relaxed to fast‑track construction. Demand is rising faster than our ability to manage its environmental footprint. https://www.legislation.gov.uk/en/uksi/2026/13
In short, the UK is heading towards a future with less water, more drought and greater pressure on already stretched resources.
The Human Cost We Don’t Want to Repeat
Technology has always reshaped work, but the speed and scope of AI feels different. It’s not just automating repetitive tasks; it’s reaching into creative, analytical and professional roles once considered “safe.”
We’ve seen what happens when industries collapse without sustainable transition plans.
When coal mining was dismantled in the 1980s, jobs disappeared but so did identity, purpose and entire local economies. The damage wasn’t just economic; it was generational.
We cannot afford to repeat that.
If AI reshapes or replaces certain roles and it will, then resilience planning cannot be an afterthought. Communities need support before disruption hits, not after.
Policy: Running Behind the Train
Governments are trying to keep up, but the pace of AI growth is outstripping policy development.
The UK has taken a principle‑based, pro‑innovation approach. It sounds flexible, but flexibility can leave gaps:
- Workforce protection is limited and non‑binding.
- Environmental oversight is fragmented and mostly voluntary.
- Transparency requirements are inconsistent.
- Transition planning is almost entirely absent.
- What happens if AI gets it wrong.
Meanwhile, the EU and OECD are moving faster on worker‑impact assessments and transparency, but global alignment is still patchy.
In short: policy is lagging behind the technology it’s meant to govern.
So What Do We Do With All This?
AI isn’t something we can roll back (not should we). It has enormous potential to improve healthcare, accelerate research, expand accessibility and help solve complex problems.
But potential doesn’t erase responsibility.
We still have choices, meaningful ones about how AI develops and who benefits from it.
That means choosing to:
- ask better questions,
- demand transparent reporting,
- set clear carbon‑accounting standards for AI,
- require sustainability and environmental impact goals across the sector,
- push for fair, enforceable policy,
- set accountability frameworks for misuse and criminal use of AI,
- protect people, professionals and communities whose skills are being displaced,
- design AI that aligns with our values, not just our ambitions.
A Call to Pay Attention
This is our chance to shape the future before it shapes us.
In the three years since I was first introduced to ChatGPT, I’ve seen for myself how convenient and effective AI tools can be, from Copilot to curated music playlists. I wouldn’t call myself an expert, just one of millions now using generative AI in everyday life.
The question isn’t whether AI is good or bad; its innovative credentials are extensive and extraordinary. The real question is whether we’re willing to build it in a way that protects people, communities and the environment.
Are we prepared to learn from the rise of social media, a sector where safeguarding arrived far too late, long after the technology had already reshaped society?
If we start connecting the dots now while we still have the opportunity, we can build an AI‑enabled future that lifts people up rather than leaving them behind.
None of this is anti‑technology. It’s pro‑people and pro‑planet. It’s simply time we did an honest accounting of the real cost of AI.
Thank you for reading.
meshikiyang.com 
Leave a comment