For the first time in seven scans, optimistic AI narratives are outrunning dystopian ones by a double-digit margin. But look at what's driving that gap and you'll want to ask harder questions.
Right now, the scary story is winning. This week, more of the news pointed to the Terminator future — the one where AI ends up in charge and ordinary people pay the price, mostly by losing their jobs. The happier Futurama future — where AI stays a helpful tool and humans keep control — slipped back. And that's the future most of us would actually pick. So this gap isn't just a number on a chart. Think of it as a scoreboard for one big question: which ending is the world heading for — the scary one, or the good one?
Seven scans in, one number keeps doing the most analytical work: the Dystopia–Futurama gap. It opened at +5 in April, spiked to +24 a week later – the series high-water mark for pessimism – and has been on an uneven retreat ever since. This week it lands at –11, the most negative reading in the entire series. Futurama is no longer chasing Dystopia. It has lapped it.
But don't pop anything yet. The question isn't whether optimistic framing is rising. It obviously is. The question is what kind of optimism, and what it is crowding out.
At 40% of this week's sources, Futurama posts its highest single-scan share since we started tracking. The signal is real: governance frameworks from the EU AI Act to Colorado's disclosure rules, alignment research breakthroughs in Constitutional AI and mechanistic interpretability, enterprise productivity reports from Deloitte and Orange Business, the MIT small-model efficiency story. These aren't puff pieces. They represent genuine institutional effort to route AI toward broad benefit.
Here's the contradiction, though. Several of the same enterprise sources driving Futurama upward are directly contradicted by sources I've filed under Fizzle. Writer's enterprise adoption survey finds 79% of organisations facing significant challenges. Terminal X puts the project abandonment rate at 42%. IBM Think reports only 25% of AI initiatives deliver expected ROI, with just 16% scaled enterprise-wide. NVIDIA is telling one story. The people actually deploying the technology are quietly telling another. When Futurama's share rises this sharply, partly on the back of vendor-commissioned surveys – NVIDIA, PwC, Deloitte – it's worth flagging the source mix.
At 29%, Dystopia dips one point from last week's 30% and two points from its recent plateau. That's not a trend. That's noise around a floor. And the floor is set by something structurally difficult to shift: 142,000 tech job cuts in 2026 with nearly half explicitly attributed to AI, as tracked by Tom's Hardware and TechTimes. Profitable hyperscalers cutting workers to fund $700 billion in AI infrastructure. The Fed flagging AI as a systemic risk. Florida pursuing Sam Altman personally over concealed harm risks. PwC's own performance study finding three-quarters of AI's economic gains captured by just 20% of companies.
Notice what that last finding does to the Futurama narrative. The same consulting firm producing optimistic AI productivity reports is also producing the data showing those gains are dramatically concentrated. Both things are true. The discourse just rarely runs them side by side.
Singularia rises from 13% to 16% – back to its scan 5 level, driven by Anthropic's recursive self-improvement research and a convergence of AGI-horizon pieces treating 2030–2035 arrival as the relevant planning assumption, not science fiction. The AGI-26 conference, TechTarget's 'so close yet so far' framing, and the International Banker's 'humans lost the monopoly on smart' piece all treat AGI as a destination question, not an existence question. That's a meaningful shift in editorial baseline.
Paperclipalypse, meanwhile, nearly vanishes – one source, 2%, a series low. The sole entry comes from Inside Global Tech's coverage of the International AI Safety Report, which uniquely highlighted criminal and state-actor weaponisation of AI capabilities. Everything else in the safety coverage this week was filed under Futurama – alignment as solvable engineering problem, governance as manageable policy challenge. The almost total absence of indifferent-system risk from this week's discourse is itself a data point. When safety conversations are optimistic about alignment, who is minding the unintended-consequence flank?
Zoom out across all seven scans and the story is this. We started in a world where nearly a third of AI discourse was dystopian and existential risk was everywhere. Over twelve weeks, the centre of gravity has migrated. Governance frameworks have multiplied, enterprise adoption has normalised, and the dominant frame has shifted from 'what could go wrong' to 'how do we operationalise this.' Dystopia hasn't collapsed – the job displacement data alone keeps its floor high – but it is increasingly a parallel track rather than the headline register.
What has structurally changed is the absence of serious reckoning with distribution. The enterprise productivity story and the 'AI cut my workforce' story coexist in the same week's scan without much friction in the wider discourse. Futurama's series high and Dystopia's stubborn floor aren't contradictions the media is resolving. They're contradictions the media is running simultaneously and hoping readers don't triangulate.
If the Dystopia–Futurama gap keeps widening in Futurama's favour while job displacement data keeps compounding and ROI concentration keeps narrowing to 20% of firms – what exactly is the optimism measuring? And who commissioned the surveys it's built on?
The articles most responsible for bending the five worlds this week.
This piece puts the starkest number on the structural story underneath this week's scan: companies that are already profitable are cutting workers specifically to redirect capital toward AI infrastructure, with entry-level engineering employment down 20%. It's the single source most responsible for keeping Dystopia's floor intact even as Futurama surges – and it reframes the productivity narrative as a transfer, not a tide that lifts all boats.
The most analytically uncomfortable source in this week's scan, because PwC appears on both sides of the ledger: optimistic enterprise predictions under Futurama and this concentration finding under Dystopia. The 74%-to-20% split is the empirical backbone of why Dystopia won't collapse regardless of how many governance frameworks proliferate – and it directly undermines the 'broadly beneficial tool' framing that inflated Futurama's count.
MIT's efficient small-model research is the kind of structural Futurama signal that isn't vendor-commissioned and isn't governance boilerplate – it points to a genuine mechanism by which AI capability could democratise rather than concentrate. In a week where much of the Futurama count rests on enterprise surveys with obvious incentives, this is the source that most legitimately earns its classification and pulls the gap in Futurama's direction.
Anthropic's own research arm documenting Claude agents autonomously conducting AI safety research and iteratively self-improving is the week's most consequential Singularia signal – not because it proves AGI is imminent, but because the institution most associated with caution is itself now publishing on recursive self-improvement as a live research programme rather than a theoretical risk. That's a significant shift in what the safety community is treating as the relevant frontier.
With Futurama at a series high partly on the strength of enterprise productivity narratives, Writer's finding that 79% of organisations face significant adoption challenges and only 29% see meaningful ROI is the direct counter-evidence the discourse is largely ignoring. It doesn't get the media amplification of the NVIDIA or Deloitte reports, but it represents the ground-level reality for the majority of organisations actually trying to deploy AI at scale.
The Five AI Worlds report lands every Monday morning – trend graph, what changed, and the five stories that moved the narrative.
Subscribe free