A jury handed Sam Altman the keys to the AI kingdom. Meanwhile, 8,000 Meta workers got their marching orders. The gap between who wins from AI and who pays for it just got wider.
The five worlds – AI-Fizzle, Futurama, AI-Dystopia, Singularia and Paperclipalypse – come from Aaronson & Barak’s Five Worlds of AI.
This week, a lot happened that helps us see where AI is heading. A court said Elon Musk was too late to stop OpenAI from becoming a for-profit company, which means one powerful lab gets even more powerful. At the same time, Meta cut 8,000 jobs while spending billions on AI. That's the pattern we keep seeing: the big companies get bigger, and workers feel the squeeze. Our data calls this the 'Distribution Gap' – the gap between the scary world where AI's rewards pile up at the top (that's the Paperclipalypse path) and the better world where AI stays a tool that helps everyone (that's Futurama). Right now, the gap is growing, not shrinking.
Let me start with a number: 7.
That's the global Distribution Gap this week – AI-Dystopia at 30%, Futurama at 23%, seven points separating them. It doesn't sound enormous. But look at the trajectory. We started this series in April at a gap of 10, watched it compress to 2 in late April, bounce back to 15 in mid-April's peak, and now it's settled at 7, the highest it's been in three weeks and rising again.
This isn't noise. This is a direction.
The question the Distribution Gap asks is simple: as AI delivers value, who gets it? The answer this week, across 281 signals from business and trade sources in Australia, North America, Europe and APAC, was: mostly the people who were already winning.
A San Francisco jury ruled against Elon Musk in his suit against OpenAI and Sam Altman. Fifteen separate publications covered it – the most widely reported story of the week by a distance. The procedural verdict means OpenAI's path to an IPO is clear. That IPO is now weeks away, not months.
Separately, Meta began cutting 8,000 jobs – its biggest reduction in years – explicitly framed as an AI efficiency push. Nvidia posted record quarterly revenue of US$136 billion. Anthropic neared a US$30 billion funding round. SpaceX filed for an IPO with AI orbital data centres at its core.
See the pattern? The industry lens – what AI companies are doing to themselves and each other – dominated this week's narrative. Of the 281 signals tracked, 153 were industry stories. Only 159 were adoption stories: organisations actually using AI in their own operations. That ratio matters. Vendor announcements and capital raises are not evidence that AI is working. They're evidence that investors believe it will.
I keep a close eye on that distinction because it's where the hype-to-reality gap lives.
Here's something worth pausing on. Australia and NZ almost always run more optimistic than the global average – our readers' home market tends to weight Futurama higher and AI-Dystopia lower. This week, the local Distribution Gap was +1 (Dystopia 27%, Futurama 25%), compared to a rolling four-week figure of –6. That's a meaningful shift toward the global pessimistic view.
What drove it? Graduate job fears at a Senate inquiry. MinterEllison cutting its intake because of AI. WiseTech's AI redundancy dispute turning ugly. Young Australians using AI but doubting it'll help them find work. The EY cyber report retracted for AI-fabricated citations. A community backlash killing a WA data centre.
These aren't abstract global forces. They're Australian institutions, Australian workers, Australian regulators grappling with what AI actually does when it lands in a real organisation.
The good news: CBA appointing its first chief AI scientist, the Australian Attorney-General addressing AI and copyright reform, Tower Insurance saving 550 days of call time with Amazon Connect, and TechnologyOne growing ARR toward $600 million partly on the back of AI bets. Australia has genuine Futurama signals. They're just not winning the week.
North America this week ran a Distribution Gap of 19 – Dystopia at 34%, Futurama at just 15%. That's the highest regional gap we've recorded in this series. The rolling four-week figure is 13, so this week is even hotter than the trend.
Meta layoffs. OpenAI IPO. Nvidia earnings. Anthropic funding. Trump executive order infighting. Workers faking AI skills out of job-loss fear. A Nobel economist questioning whether the productivity case holds up. Goldman Sachs finding little evidence of AI-driven labour market reshuffling – which, paradoxically, is a Fizzle signal as much as a Dystopia one.
North America is telling a story about concentrated power and uneven rewards. Every major signal this week pointed the same way.
Europe is the contrarian this week. Distribution Gap of –2, meaning Futurama (30%) actually beat Dystopia (28%) – the only region where that happened. Malta rolled out free premium AI access for all citizens tied to literacy training. Legora raised €600 million with Nvidia backing. Legaltech startups are proliferating, and the EU AI Act is being actively debated. Greenpixie raised £4.7 million to cut AI energy waste for Fortune 1000 firms.
Europe's rolling four-week Gap is +9, so one optimistic week doesn't reverse the trend. But it's a real signal: European AI policy, for all its friction, is producing at least some stories about spreading access rather than concentrating it.
APAC (Singapore and Hong Kong) held its Distribution Gap at 11 – identical to its rolling four-week average. Stability, but in the pessimistic direction. China's three state-owned telcos racing into AI token subscriptions. Samsung workers in dispute over bonus gaps driven by AI chip dynamics. AI giants concentrating hiring in Singapore. DeepSeek recruiting elite quant talent to build AI agents.
APAC knows who controls AI infrastructure in its region. The signals this week didn't suggest that's changing.
This week's driver breakdown tells a clear story. Economy (86 signals) and technology (122) dominated, with regulation (43) and talent (44) not far behind. Trust came in at 48.
AI-Dystopia: almost entirely economy and talent – mass layoffs, capital concentration, workforce displacement. The Musk-OpenAI verdict added a regulation signal, but even that was fundamentally about who controls and benefits from a dominant lab.
Futurama: technology and talent – new roles emerging, AI tools delivering measurable value, HR leading adoption thoughtfully. Regulation played a small positive role (Malta, Singapore, Australian government procurement with human oversight).
Singularia: technology almost entirely. Google I/O, Gemini 3.5 Flash, SAP's autonomous enterprise vision, agentic cloud infrastructure. Labs becoming agent labs. The industry has made a collective bet on autonomy as the next frontier.
Paperclipalypse: trust, overwhelmingly. AI-accelerated attacks. EY retracting a hallucinated report. AI chatbots leaking phone numbers. Deepfake fraud beating enterprise workflows. Shadow AI agents that agencies can't audit. The harms are real and the containment is lagging.
AI-Fizzle: technology and economy, equally. Australian AI ambition colliding with reality. SAP's Joule adoption described as 'minimal'. 43% of enterprise AI initiatives forecast to fail. API bills exceeding revenue for simple tasks. The pilot-to-production gap diagnosed as the defining enterprise failure mode.
The structurally absent story this week: small and medium business adoption. SMEs appear almost only as victims – of Google's AI Search overhaul killing organic traffic, of not being data-ready, of fearing AI attacks they can't afford to defend against. We have almost no signals of SMEs actually using AI and winning. That's a genuine blind spot in the discourse, and it matters enormously for Australia and NZ, where SMEs are the economy.
Also missing: workforce stories with real upside that aren't vendor-sourced. The Futurama signals in the HR space this week were almost entirely from HR publications quoting HR software vendors. Independent evidence that AI is genuinely improving workers' lives – not just their employers' productivity metrics – was thin.
If you're a CEO or CIO watching this data, here's the plain version: the industry is telling you AI is working. The adoption data is telling you it often isn't – yet. And the workforce data is telling you your people are scared.
The Distribution Gap is the number to watch. When it's positive and rising, value is concentrating upward. Your job is to make deliberate choices about where your organisation sits on that spectrum. Are you cutting costs with AI and pocketing the savings? Or are you sharing the upside with the people whose work AI is augmenting?
That question isn't just ethical. It's strategic. The AI-Fizzle signals this week – 51% of professionals saying AI tools lower their productivity, AI spending disconnected from measurable value, pilot projects dying before production – all trace back to organisations that haven't answered it honestly.
Pick a specific workflow. Measure the value AI creates in it. Then decide, explicitly, where that value goes. That's not a soft question. That's the Distribution Gap, inside your organisation, waiting for your answer.
The articles most responsible for bending the five worlds this week.
A jury said Elon Musk waited too long to sue, so OpenAI and Sam Altman won. This clears the way for OpenAI to become a for-profit company and go public. It's big because it means the world's most powerful AI lab just got a green light to grow even more powerful, with very little standing in its way.
Meta is cutting 8,000 jobs and spending the savings on AI. The company says it will be more efficient, but the people who lost their jobs aren't sharing in that efficiency. This story landed in NZ and Australian outlets because it shows the pattern that worries workers everywhere: AI's wins go to the company, not the workforce.
Australian students told a Senate inquiry that AI is taking away the entry-level jobs they were training for, and that it's stopping them from learning real skills. This is a local, on-the-record signal – not a survey, not a vendor prediction. Young Australians are telling parliament they feel locked out of the future AI is building.
EY published a cyber security report full of made-up citations and wrong information, then had to pull it back. This is one of the biggest professional services firms in the world. It shows that AI can produce confident-sounding nonsense, and that even expert organisations can miss it until it's too late.
Australian business leaders are excited about AI but struggling to make it actually work in their organisations. The gap between big plans and real results is growing. This matters because it's not a technology problem – it's a leadership and execution problem, and it's happening right here at home.
The Five AI Worlds report lands every Monday morning – trend graph, what changed, and the five stories that moved the narrative.
Subscribe free