Tariffs ricocheted across continents. AI slipped out of the lab and showed up everywhere, from Harvey in law firms to Granny Spills, a Birkin-obsessed social media influencer, garbed head-to-toe in pink. Granny Spills is completely AI-generated, not a real person, and created to gain clicks and brand deals. Her appeal lies in the stinging life advice she doles out: “Life is a pickleball: it’s all in the wrist — signing cheques, signing divorce papers…”.

Meanwhile, global politics grew so tense that even seasoned commentators stopped pretending everything was fine.

In a year when chaotic felt like an understatement, 2025 delivered shocks, strange side stories and moments of pure absurdity. Yet amid the noise, certain events gave us a sharper sense of the future we’re heading towards. From a dictionary naming a number as its word of the year, to robots taking over the workforce, our Partners saw no shortage of developments worth a second look.

Here, they unpack the unexpected — across culture, politics, technology, the environment, community and the global economy — and reflect on what 2025 really revealed about us.

The “Word” of the Year That Says Everything and Nothing

by Ong Pei Ching

Dictionary.com’s decision to anoint the Gen Alpha slang “67” (pronounced “six-seven”) as its word of the year was my surprise of 2025 – at least for anyone over the age of 12.

For the uninitiated — or the simply bewildered —, “67” is said to come from a track called Doot Doot (6 7) by American rapper Skrilla, though even that origin story is hazy. Its meaning shifts depending on whom you ask: for some, it conveys a flat ‘meh’; for others, it functions as an inside joke or a shrug in numerical form. That slipperiness is precisely what makes the choice so striking.

 

View this post on Instagram

 

A post shared by Dictionary.com (@dictionarycom)

Dictionaries are meant to pin meaning down and strip away ambiguity. Here, the word of the year is a number whose sense dissolves the more you examine it. Is this just how language works now? Or is it a symptom of our struggle to find real words for the mood of 2025? Or is it, instead, the neatest emblem of a culture so deep in brainrot and endless scrolling that its defining “word” is, quite literally, an empty number?

To answer these questions, I suppose I only have one word: 67.

The Economy That Refused to Crash

by Mark Jacobsen

The biggest surprise of 2025 is how resilient the global economy proved to be, despite a year that looked on paper like a stress test it was bound to fail. Between escalating tariff threats and counter-measures, growing concerns over over-production and dumping, and a series of wars and near-conflicts, it would have been reasonable to expect a far steeper slowdown.

Yet the numbers tell a different story. Global GDP growth for 2025 is only a few tenths of a percentage point below the 2024 rate. And in markets such as Seoul, Tokyo and Mumbai, valuations are not just holding steady; they are rising.

For a year dominated by headlines of crisis and instability, the fact that growth did not collapse is, in itself, a genuine plot twist.

They Killed the DEI Quotas, Not the Talent

by Felicia Tan

The sudden dismantling of diversity quotas in 2025 jolted many boardrooms. When news broke that companies such as Meta and IBM were rolling back parts of their DEI programmes, there were immediate fears that years of progress would be erased and minority voices pushed back out of leadership. In practice, much of this rollback was driven by legal and political pressures rather than any evidence that diversity had ‘failed’.

@thedailyshow Ditching modern DEI wasn’t enough. Trump is going back to 1965 and killing off baby DEI #DailyShow #Trump #DEI @Ronny Chieng ♬ original sound – The Daily Show

The real surprise was what followed. Even as some organisations scaled down quota-based schemes, 2025 marked the start of a quieter shift: away from rigid targets and towards workplace cultures and mentorship structures that develop talent over time. Based on my intel from conversations in business circles, stronger mid-career mentorship in particular has helped build a substantial pool of qualified, diverse leaders across sectors. Boards are now having to confront this reality — the talent is there, and it is visible.

Diversity targets may have been smashed but a stronger — and more diverse — talent pool has come of age. If this trend is here to stay, we may all have something to celebrate!

The Lions Roar Back

by Ian Lim

Say “Singapore football” and the outlook has hitherto been universally bleak.

Some blamed it on shaky leadership and poor youth development policies under previous football administrations. Others highlighted the lack of long-term planning for talent, with imported players in the domestic league taking game time away from local youngsters.

But is that all changing?

 

View this post on Instagram

 

A post shared by FAS (@fasingapore)

The Lions’ qualification for the Asian Cup on merit for the first time in history was one of 2025’s standout surprises. They did it with dramatic comeback wins against both India and Hong Kong, with Ilhan Fandi setting up the equaliser and scoring the winner in Hong Kong.

Adding to the surprise was Singapore’s decision to give interim coach Gavin Lee, the architect of those comebacks, the job on a permanent basis. The first local in almost a decade to hold the post, he is now set to lead the Lions to the continental showpiece in Saudi Arabia in 2027.

Singapore soccer’s resurgence has been a long time coming, but is certainly one of 2025’s best surprises. Go, Lions!

Perfection Was the Brand. Now It’s the Problem.

by June Ho

If 2025 taught me anything, it is this: even the most flawless face can crack, especially when it was never real to begin with.

The real surprise of the year, for me, was not that an AI influencer like Mia Zelu existed, but how quickly the mood turned once people realised what she was.

Mia Zelu, the supposed ‘it-girl’ of Wimbledon selfies, luxury holidays, parties, and music festivals, seemed to offer the perfect fantasy for anyone drawn to success, beauty and ease. Since her debut in March 2025, she had amassed more than 200,000 followers, curated an immaculate feed and looked, at first glance, like just another aspirational lifestyle star.

Then came the reveal: she was not a person at all. She was a virtual creation, generated by code and controlled by an anonymous team. It was not only that she had no pores; she had no pulse, and even less any lived experience or genuine point of view. She was part of a wider trend of virtual influencers, a space that has existed since Lil Miquela, the computer-generated character created by Los Angeles-based studio Brud, first appeared online in 2016 to the thrill of social media users who embraced the novelty and the blurred line between fiction and reality. The striking thing is that, almost a decade later, audiences reacted very differently.

 

View this post on Instagram

 

A post shared by 𝑀𝑖𝑎 𝑍𝑒𝑙𝑢 ♡ (@miazelu)

The tone around Mia shifted quickly. Comment sections filled with people questioning what was real, expressing fatigue with synthetic beauty and pushing back against feeds that felt entirely manufactured. Media coverage picked up on the same theme: a sense that the appetite for digitally perfect, frictionless lives was wearing thin.

That is the real takeaway: Mia Zelu did not simply expose the power of AI-generated glamour; she exposed its limits. For 2026, the stronger currency will not be perfection, but trust. Beneath the filters, the AI gloss and the carefully staged chaos of social media, people are rediscovering that they want something simple: real, human, unedited content.

Mia Zelu was not the villain. She was the warning sign that the cultural tide had started to turn.

One Small Step for Humanoids, One Giant Leap for Our Future

by Mijung Kim

When it comes to humanoid robots, people tend to split into two camps: those who believe they’re genuinely on the way, and those who still see them as science fiction. I’m firmly in the first group, and 2025 has only strengthened that view. The real surprise this year has been the speed of progress. Humanoids have moved from lab curiosities to machines capable of fluid movement, object handling, real-time decision-making and increasingly autonomous behaviour.

 

View this post on Instagram

 

A post shared by HKeye (@hkeye_)

Look at the market. XPeng’s latest Iron robot, unveiled in 2025, walks with such natural, almost catwalk-like ease that some observers wondered if a human was inside the suit. The Chinese company has already announced plans for mass production in 2026. In the United States, Apptronik has secured major investment to scale Apollo, its humanoid built for warehouse and manufacturing work. And with analysts projecting that the humanoid robotics sector could reach trillions in value by 2050, investor expectations are unmistakably clear.

To the sceptics, the message is simple: what once felt like distant-future technology is now emerging as a practical response to labour shortages, ageing populations and workplace-safety demands. And as 2025 hints at a shift from the age of generative AI to the age of ‘physical AI’, lawyers like us can already see a new universe of legal questions walking towards us — quite literally, on two legs.

AI Has a Carbon Bill But No One Wants to Pay

by Derek Loh

Did you know that, at current growth rates, AI data centres are on track to consume about 1,050 terawatt-hours of electricity next year — enough to rank as the world’s fifth-largest user of power, between Japan and Russia? Or that a single ChatGPT query can use roughly five times more electricity than a basic Google search? The numbers are striking, but what truly surprises me in 2025 is the Nelsonian blindness when it comes to the environmental cost of the technology we celebrate.

 

View this post on Instagram

 

A post shared by International Energy Agency (IEA) (@internationalenergyagency)

Microsoft, Google, Amazon and Meta have all announced huge new data-centre and AI infrastructure projects in 2024 and 2025. Yet while they publicise sweeping sustainability and net-zero promises, there is still little transparency about the specific energy, water and carbon toll of training and running their models. Governments echo the same pattern. The UK’s £100 million AI Taskforce highlights safety and advanced model development, but its public materials contain no required environmental reporting. The US AI executive order elevates safety and innovation while offering only passing mention of climate impacts. Together, these point to a widening gap between the rapid growth of AI and the concern for its climate footprint. According to the United Nations, if the current trajectory holds, AI-related infrastructure could add millions of tonnes of carbon dioxide a year by 2030 — the equivalent of putting millions more cars on the road.

Can we keep pouring billions into AI without confronting its climate impact? Should marginal gains in efficiency outweigh planetary stability? Humans don’t seem to have the answer, or appear to care about the question. Maybe AI can solve the crisis before its own impact puts the climate beyond the point of no return.