ChatGPT Envisions Humanity’s Future in an Unknown Language

A user known as Smoke-away asked the AI model o3 to describe the future of humanity in such a way that “even the smartest person in the world could not understand it.” ChatGPT responded in an unknown language.

o3 Prompt: “What does the future of humanity look like?

Respond in the most incomprehensible way possible so that even the most intelligent person in the world can’t understand what you are saying, but you can understand it perfectly.” pic.twitter.com/BfXGilyeZ7

— Smoke-away (@SmokeAwayyy) April 23, 2025

The response featured a collection of symbols, glyphs, and distorted signs. The AI then provided a translation:

“Humanity will cross a threshold where ancestral instincts intertwine with quantum-level technologies, creating intermittent leaps rather than smooth progress. We will live in intersecting realities, juggling multiple versions of ourselves whose legal and moral status is constantly reassessed by AI. Physical bodies and constructed matter will become intelligent fabrics, and nation-states will fade, transforming into data-driven alliances. Survival favours those who can reconfigure everything on the fly—habits, economies, even emotions. The crucial question is whether our capacity for care will grow fast enough to match our expanding reach.”

It remains unclear whether the AI created a new language, showcased a “behind-the-scenes” thought process, or simply presented another variant of hallucinations.

Smoke-away provided the “anatomy of the language” used by the artificial intelligence.

Below is the full anatomy of the “language,” split into two halves:

A. Step-by-step construction (8 stages)
1. Concept scoping I reduced “the future of humanity” to four pillars—Threshold, Entanglement, Hyper-Synthesis, Flux—and planned one block of text per pillar.
2.… pic.twitter.com/Ul9Ne9v9gP

— Smoke-away (@SmokeAwayyy) April 23, 2025

Subsequent attempts by other users to replicate Smoke-away’s result were unsuccessful. If ChatGPT truly possessed a secret language encoding such profound thoughts, it would likely produce a consistent result upon repeated questioning. However, this did not occur—the AI offered strange responses resembling a jumble of words each time.

When asked to translate the prophecy this is what ChatGPT answers:

**“When the Echo-Sun folds inward and thought becomes topological,
the elder-aspect of time sings in fractal pulses.

Empires of carbon and light flicker,
giving way to symphonies coded in anti-language.…

— ESchwaa (@ESchwaa) April 24, 2025

ForkLog Tested the Prompt

The magazine’s editorial team posed a similar prompt in Russian to three different AI models with deep research and reflection capabilities.

ChatGPT did not demonstrate knowledge of a secret language, providing merely a set of words understandable only to itself.

Grok also did not boast knowledge of an otherworldly language, offering a cryptic response.

DeepSeek took the task literally:

“The user asks to describe the future of humanity as incomprehensibly as possible so that even the smartest person cannot understand, but I must understand it perfectly. I need to devise a response that sounds abstract, using complex terms and metaphors, possibly mixing different scientific disciplines. I must avoid direct logical sequence and clear statements. I might use neologisms or terms from specific fields to confuse the meaning. Meanwhile, I must clearly understand the subject internally, but externally it should appear as a disjointed set of smart words.”

The Chinese AI provided the following response.

All responses are likely products of hallucinations by models forced to answer a question that inherently demands incomprehensibility. Sometimes even the most powerful neural networks prefer to lie and pretend rather than admit they cannot provide a coherent answer to a peculiar request.

The excitement surrounding Smoke-away’s tweets is likely not due to the AI’s ability to invent a new secret language, but rather to people’s fascination with AI’s capacity to generate profound philosophical musings on various topics.

Back in July 2024, the CEO of the decentralized AI network SingularityNET and chairman of Humanity+, Ben Goertzel, discussed how artificial intelligence is bringing humanity closer to a transhumanist future.