Zhipu AI founder says we are on a clear path to AGI with AI beginning…

What was claimed

Zhipu AI founder says we are on a clear path to AGI with AI beginning to learn the 'self', achieving self-awareness, emotion, consciousness, and eventually ASI via recursive self-improvement where AI trains AI

Our verdict

Needs Caution

Tang Jie and Zhipu leadership do talk explicitly about an AGI mission and China’s path to AGI, and Zhang Peng has said things like “2024 will be the first year of AGI.” However, describing this as a “clear path” overstates the certainty; their language is aspirational and visionary, not a claim that the path is clearly established and guaranteed. Tang Jie’s letter explicitly states that after learning what ‘I’ is, AI will “further on… touch human emotion,” indicating he envisions AI systems interacting with or modeling emotion. The AI‑generated paraphrase that AI will “achieve emotion” is stronger than his original phrasing, which is future‑oriented and somewhat vague rather than a concrete claim of full emotional experience.

2 of 3 AI systems agree8 sources citedChecked Jul 12, 2026

Check your own claim

Paste any statement, headline, or AI answer — 3 independent AIs verify it in seconds, with sources.

Key findings

The Zhipu AI founder says we are on a clear path to AGI.

Misleading78%

1 of 2 AIs agree·ChatGPT: Can’t verify

Current AI is beginning to learn the 'self' in a way that implies genuine self-awareness, emotion, and consciousness

Can’t verify70%

3 of 6 AIs agree·ChatGPT: Misleading, Perplexity: Misleading, Perplexity: Incorrect

Eventually ASI will arise via recursive self‑improvement where AI trains AI.

Can’t verify75%

1 AI checked

Verified against 8 sources

Detailed Analysis

Parts of the statement accurately reflect what Zhipu co‑founder Tang Jie has written about AI learning the concept of ‘I’ and eventually touching emotion and consciousness. Other parts, like claiming we are on a clear path to AGI and that current AI is beginning to learn the self in a way that implies real self‑awareness, emotion, and consciousness, go beyond what can be verified and mix speculation with fact. The recursive self‑improvement/ASI part is also speculative and not grounded in current capabilities.

Why this verdict

Parts of the statement accurately reflect what Zhipu co‑founder Tang Jie has written about AI learning the concept of ‘I’ and eventually touching emotion and consciousness.
Other parts, like claiming we are on a clear path to AGI and that current AI is beginning to learn the self in a way that implies real self‑awareness, emotion, and consciousness, go beyond what can be verified and mix speculation with fact.
The recursive self‑improvement/ASI part is also speculative and not grounded in current capabilities.

Claims checked

Current AI is beginning to learn the 'self' in a way that implies genuine self-awareness, emotion, and consciousness

Incorrect90%

Tang Jie talks about future AI beginning to learn what ‘I’ is and eventually touching emotion and consciousness, but this is aspirational and forward‑looking.[7] Contemporary expert sources state that today’s AI systems are not self‑aware or conscious, and only simulate introspection without genuine subjective awareness or emotion.[3] [5]

This claim may mislead readers without added context.

Zhipu AI founder says we are on a clear path to AGI

Misleading78%

Tang Jie and Zhipu leadership do talk explicitly about an AGI mission and China’s path to AGI, and Zhang Peng has said things like “2024 will be the first year of AGI.”[1] [7] However, describing this as a “clear path” overstates the certainty; their language is aspirational and visionary, not a claim that the path is clearly established and guaranteed.[1] [7]

This claim may mislead readers without added context.

Zhipu AI founder says AI will begin to learn the 'self', achieving self-awareness

Misleading83%

In his letter, Tang Jie writes that “AI will begin to learn what ‘I’ is — what self‑cognition means,” which supports the idea that he expects AI systems to model a notion of self.[7] However, he does not clearly claim that present systems have already achieved genuine self‑awareness, and broader expert commentary stresses that current AI is not self‑aware but only mimics introspective language.[3] [5]

This claim may mislead readers without added context.

What other AI think

Zhipu AI founder Tang Jie has stated that we are on a clear path to AGI, and he argues that AI will begin to learn what the "self" is and what self-awareness means, may touch human emotion, and consciousness itself. Models are already participating in training future models, and Tang appears to believe this represents the beginning of recursive self-improvement. The claim accurately reflects his recent statements.</reasoning> <parameter name="claims">[ { "claim": "Zhipu AI founder says we are on a clear path to AGI", "assessment": "TRUE", "confidence": 0.95, "explanation": "Tang Jie, Zhipu AI founder, has stated we are on a clear path to AGI. This is documented in his recent internal letter." }, { "claim": "AI beginning to learn the 'self'", "assessment": "TRUE", "confidence": 0.95, "explanation": "Tang argues that AI will begin to learn what the 'self' is and what self-awareness means. This directly matches the claim." }, { "claim": "Achieving self-awareness, emotion, consciousness", "assessment": "TRUE", "confidence": 0.95, "explanation": "Tang states AI may touch human emotion and consciousness itself. The claim accurately reflects his progression of capabilities." }, { "claim": "Eventually ASI via recursive self-improvement where AI trains AI", "assessment": "TRUE", "confidence": 0.90, "explanation": "Models are already participating in training future models, and Tang appears to believe this represents the beginning of recursive self-improvement. His essay describes progression to self-evolving intelligence and eventually ASI via AGI." } ]

Why this verdict

Zhipu AI founder Tang Jie has stated that we are on a clear path to AGI, and he argues that AI will begin to learn what the "self" is and what self-awareness means, may touch human emotion, and consciousness itself.
Models are already participating in training future models, and Tang appears to believe this represents the beginning of recursive self-improvement.
The claim accurately reflects his recent statements.</reasoning> <parameter name="claims">[ { "claim": "Zhipu AI founder says we are on a clear path to AGI", "assessment": "TRUE", "confidence": 0.95, "explanation": "Tang Jie, Zhipu AI founder, has stated we are on a clear path to AGI.
This is documented in his recent internal letter." }, { "claim": "AI beginning to learn the 'self'", "assessment": "TRUE", "confidence": 0.95, "explanation": "Tang argues that AI will begin to learn what the 'self' is and what self-awareness means.

The statement mixes an attribution to a Zhipu founder with broad, time‑sensitive technical claims. Attribution and exact wording are not clearly supported by reliable primary sources. The technical assertions (AI already having self, emotion, consciousness or inevitable ASI via recursive self‑improvement) are speculative and not established facts.

Why this verdict

The statement mixes an attribution to a Zhipu founder with broad, time‑sensitive technical claims.
Attribution and exact wording are not clearly supported by reliable primary sources.
The technical assertions (AI already having self, emotion, consciousness or inevitable ASI via recursive self‑improvement) are speculative and not established facts.

Claims checked

AI is beginning to learn the 'self', achieving self‑awareness, emotion, and consciousness.

Misleading90%

Researchers and commentators discuss emergent self‑modeling and introspection in models, but there is no consensus or reproducible evidence that deployed AIs have genuine self‑awareness, subjective emotions, or consciousness.

This claim may mislead readers without added context.

The Zhipu AI founder says we are on a clear path to AGI.

Can’t verify65%

I found reporting and social posts implying Zhipu founder Tang Jie has discussed long‑term AGI goals and AI taking responsibility, but I could not locate a reliable primary source with the exact quoted wording 'clear path to AGI.'

Eventually ASI will arise via recursive self‑improvement where AI trains AI.

Can’t verify75%

Recursive self‑improvement is a widely discussed hypothesis and some experts predict it as a possible path, but there is no verified timeline or proof that ASI will inevitably emerge this way.

Share this result

Try DoubleChecks