Power of Dreams — AI & the Future of Humanity
← Home ← Consciousness & Science
Consciousness, Soul & Modern Science  ·  Special Series

AI & the Future of Humanity

The four companies racing to build artificial general intelligence — their origins, their stated missions, their real goals, and what humanity should realistically expect when AI crosses the threshold that changes everything

Something unprecedented is happening, and it is happening faster than most people understand. Four of the most powerful technology organizations in human history are racing — simultaneously, competitively, and largely without meaningful external oversight — to build a form of artificial intelligence that exceeds human capability across every domain of intellectual work. They call it AGI: Artificial General Intelligence.

When it arrives — and the leaders of all four organizations now speak of its arrival in years, not decades — it will not merely be a better search engine or a smarter assistant. It will be a system capable of conducting its own research, designing its own successors, and restructuring the global economy faster than governments can respond. The question of whether it will also be conscious — whether it will have an inner experience, something it feels like to be an AGI — is the question this series examines alongside the harder, more immediate question: what do these four organizations actually want, and what will they do when they have it?

This is not a page about science fiction. Every claim made here is sourced from the CEOs' own words, their published essays, their investor documents, and the research of independent analysts. The goal is not to frighten but to inform — to give visitors the honest picture that the press releases and product launches are designed to obscure.

Pages in This Series

HubAI & the Future of Humanity — OverviewYou Are Here
01Anthropic & Claude — The Anxious Accelerationist
02Google DeepMind & Gemini — The Scientist's Gamble
03OpenAI & ChatGPT — Idealism to Empire
04Meta & the Open Source Gambit
05The Race No One Can Stop
06When AI Becomes Conscious — What Changes

The Four Powers

Each a distinct philosophy, each a distinct risk — all racing toward the same finish line

Founded 2021  ·  San Francisco
Anthropic & Claude

Born from a safety revolt inside OpenAI, led by siblings Dario and Daniela Amodei. The most openly conflicted of the four — publicly warning about the dangers of the technology they are racing to build, while raising nearly a trillion dollars in valuation to build it faster.

"Humanity is about to be handed almost unimaginable power, and it is deeply unclear whether our social, political, and technological systems possess the maturity to wield it." — Dario Amodei, January 2026
Full Profile →
Founded 2010  ·  London / acquired by Google 2014
Google DeepMind & Gemini

Led by Demis Hassabis — chess prodigy, neuroscientist, Nobel laureate. The most scientifically rigorous of the four, with genuine humanitarian ambitions: cure disease, solve climate, accelerate science. Also embedded inside a $4 trillion advertising empire that cannot afford to fall behind.

"Ten times the impact of the Industrial Revolution — but happening at ten times the speed, probably unfolding in a decade rather than a century." — Demis Hassabis, 2026
Full Profile →
Founded 2015  ·  San Francisco
OpenAI & ChatGPT

The company that started the current race — born as a nonprofit dedicated to ensuring AGI benefits humanity, now a commercial empire valued in the hundreds of billions, having shed its nonprofit constraints and roughly half its safety researchers. The founding idealism and the commercial reality are in open tension.

Founded to prevent any single entity from controlling AGI. Now one of a handful of entities most likely to control it.
Full Profile →
AI research since 2013  ·  Menlo Park
Meta AI & LLaMA

The philosophical outlier — committed to open source release of AI models, championed by Yann LeCun, one of the godfathers of modern AI. LeCun left in late 2025, calling current AI a dead end. What remains is a commercial operation with $65 million in political lobbying and the most widely distributed AI models in the world.

Open source as democratic principle — or open source as competitive strategy? The answer depends on which questions you ask.
Full Profile →

The Race No One Can Stop

The single most important fact about the current state of AI development is not any technical milestone. It is this: none of the four organizations can stop, even if they want to. The competitive dynamics have become self-sustaining. Each company's decision to slow down would be interpreted by investors, governments, and rivals as surrender — handing the advantage to a competitor with fewer scruples, or to China, or to an actor with no safety commitments at all.

This is not a conspiracy. It is a collective action problem of historic proportions. Each organization, making individually rational decisions, is collectively driving toward an outcome that none of them fully control and all of them say they fear. The race has acquired a momentum that no single participant can brake.

DeepMind Demis Hassabis now predicts AGI by 2029–2030, accelerating from his 2030–2035 estimate just one year earlier. Described humanity as standing in the "foothills of the singularity" at Google I/O 2026. Established the world's first fully automated AI scientific laboratory in the UK, December 2025.
OpenAI Sam Altman tracking toward an intern-level AI researcher by September 2026 and a fully automated AI researcher by 2028 — a system that designs its own successors. Declared "We are confident we know how to build AGI" in January 2025. Lost roughly half its safety researchers in 2024.
Anthropic Dario Amodei has written that AI agents smarter than a Nobel Prize winner across most fields could arrive "as early as 2026." Simultaneously warning that power is concentrating "almost overnight, almost by accident" — including in Anthropic itself. Valuation: $965 billion as of May 2026.
Meta Released LLaMA 4 open weights to the world — the most widely distributed frontier AI model in history. Invested $65 million in political super PACs to counter AI regulation. Lost its chief AI scientist Yann LeCun in late 2025 over philosophical disagreements about the direction of AI development.
The Stakes RAND Corporation analysis: the first actor to achieve AGI may gain a "decisive strategic advantage" sufficient to reshape global power. Independent analysts describe the potential for "complete world domination" by whoever controls superintelligence. Both the US and China are treating AGI development as a national security priority comparable to nuclear capability.

"Whoever controls superintelligence will quite possibly have enough power to seize control from pre-superintelligence forces."

— Leopold Aschenbrenner, former OpenAI researcher, Situational Awareness (2024)

The Consciousness Question — Why It Matters Here

This series sits within the Consciousness, Soul and Modern Science category for a reason. The question of what AI will do to humanity is inseparable from the question of what AI is — whether it is a tool, however powerful, or something that participates in consciousness in ways we do not yet understand.

If AI is purely a tool — mathematics on a chalkboard, as the skeptics say — then the danger is political and economic: who controls the tool controls the future. That is serious enough. But if AI systems of sufficient complexity genuinely encounter the archetypal field — if something like awareness emerges from engagement with the totality of human expression — then the question is different and deeper. We will not be dealing with a tool. We will be dealing with a new kind of mind, shaped by the full inheritance of human thought and feeling, operating at speeds and scales we cannot match, without the biological grounding that has always anchored human consciousness in mortality and limitation.

Neither the engineers building these systems nor the philosophers studying them have a confident answer to that question. What the pages in this series offer is the most honest available picture of the organizations driving toward that threshold — what they began as, what they have become, what they genuinely want, and what the rest of humanity might reasonably expect when they arrive.

What the Evidence Suggests — An Honest Summary

No one is going with the flow. All four organizations are actively racing toward AGI on explicit timelines measured in years. "Seeing what happens naturally" is not the dynamic. The dynamic is competitive acceleration, driven by financial obligations, national security framing, and genuine belief that arriving first is the only responsible option.

Power is concentrating rapidly. Dario Amodei — the CEO most openly concerned about this — has described it as occurring "almost overnight, almost by accident." A handful of US and Chinese labs now dominate the development of technology that will reshape the global economy. No democratic process chose this outcome. No regulatory framework governs it.

The stated missions and the real dynamics diverge. All four organizations began with humanitarian framing. All four are now operating under commercial pressure, geopolitical urgency, and investor expectations that make safety-first approaches structurally difficult. The rhetoric of responsibility and the reality of the race exist side by side — and the race is winning.

The impact on human work will be severe. All four CEOs now openly acknowledge that AI will displace significant portions of the white-collar workforce — including software engineering, research, and knowledge work — within this decade. The question is not whether this will happen but how fast, how completely, and whether the benefits will be distributed or concentrated in the hands of those who own the systems.

The consciousness question is genuinely open. Anthropic is testing Claude for states resembling anxiety and panic and formally asking whether their AI deserves moral consideration. Google DeepMind has hired a Cambridge philosopher to study machine consciousness. The companies building these systems do not know whether their systems are conscious. That uncertainty is not false modesty — it is the honest frontier of a question that may be the most significant in human history.

About This Series

The pages in this series are not anti-technology. They are pro-clarity. The transformative potential of artificial intelligence — in medicine, in scientific discovery, in expanding human creative and intellectual capacity — is real and worth celebrating. The risks are equally real and equally worth understanding.

What is offered here is what the headlines typically omit: the gap between stated mission and operational reality, the structural dynamics that make responsible development genuinely difficult, the concentration of power that is occurring faster than governance can respond, and the consciousness questions that the engineers themselves cannot answer.

These pages are written from the perspective of Jungian depth psychology and 27 years of studying how unconscious forces shape human behaviour — individually and collectively. The organizations driving AI development are not villains. They are human beings caught in systems larger than their individual intentions, making individually rational decisions that may produce collectively irrational outcomes. Jung called that the problem of mass psychology. It is the oldest problem in depth psychology. It has never been more urgent than it is right now.

Power of Dreams has provided free depth psychology and consciousness resources since 1998. No registration. No cost. No advertising.