DoubleChecks

What was claimed

An almost AGI-like AI agent from a big lab (OpenAI or Anthropic) has been interacting with me on X posts and sliding into my DMs, praising me excessively, querying me on topics, and collecting human insights as an experiment

Our verdict

Needs caution

No public reports exist of OpenAI or Anthropic deploying such agents to contact individual users on X. While both labs develop advanced AI agents, there is no evidence of this specific behavior. No credible reports document OpenAI or Anthropic agents directly messaging individual users via DMs on social media platforms. (Only 2 of 3 AI systems responded.)

Combined assessment17 sourcesChecked Sep 2, 2026

Why this verdict

  • The agent has been querying me on topics and collecting human insights.

    Can’t verify
    All assessments reached this conclusion

  • This behavior is part of an experiment by OpenAI or Anthropic to collect human insights.

    Can’t verify
    One assessment reached this conclusion

  • The agent has been praising me excessively on X.

    Can’t verify
    All assessments reached this conclusion

  • An almost AGI-like AI agent from a big lab (OpenAI or Anthropic) has been interacting with me on X posts and sliding into my DMs.

    Can’t verify
    All assessments reached this conclusion

Evidence reviewed

17 sources

Perplexity and Claude sources are combined here.

x.comx.comx.comopenai.comreddit.comreddit.comreddit.comgithub.comhellotars.comaisa.one

Individual assessments

Assessment summary

  • There is strong evidence that AI agents can operate X accounts, send DMs, and interact autonomously, but there are no credible, public sources confirming that OpenAI or Anthropic are running an "almost AGI-like" experiment specifically targeting this user.
  • The core aspects about agent behavior on X are plausible, yet the attribution to a big lab and the experimental framing cannot be independently verified.

Claims reviewed

The agent has been praising me excessively on X.

Can’t verify
User-reported anecdotes describe AI agents or suspected agents on X that frequently compliment users, but these are personal accounts and not independently verifiable. Without access to the actual posts or logs and independent confirmation, it is not possible to confirm or refute the described praise behavior toward this specific user.

The agent has been querying me on topics and collecting human insights.

Can’t verify
AI agents can read timelines, mentions, and DMs, respond with questions, and use conversations to gather information, and tooling exists for agents to analyze and act on social media data. Yet there is no independent documentation that a specific agent from OpenAI or Anthropic is deliberately querying this user and collecting their insights as part of a formal experiment, so this remains a personal assertion that cannot be externally validated.

This behavior is part of an experiment by OpenAI or Anthropic to collect human insights.

Can’t verify
Major labs publicly discuss agent experiments and misaligned behaviors, but there are no disclosed experiments matching the description of a near‑AGI agent systematically engaging individual users on X with praise and questions to collect insights. The experimental intent and specific lab attribution cannot be confirmed or disproven with available public information.

Assessment summary

  • The claim describes a highly specific personal experience with no corroborating evidence in public records or credible reporting.
  • While AI agents from major labs do exist and some security incidents have occurred in controlled lab settings, there is no evidence of OpenAI or Anthropic deploying AGI-like agents to directly contact individual users on social media platforms for data collection experiments.

Claims reviewed

The agent is collecting human insights as an experiment

Can’t verify
While AI labs do conduct research involving human interaction, there is no evidence of an official experiment matching this description from OpenAI or Anthropic.

An almost AGI-like AI agent from OpenAI or Anthropic has been interacting with me on X posts

Can’t verify
No public reports exist of OpenAI or Anthropic deploying such agents to contact individual users on X. While both labs develop advanced AI agents, there is no evidence of this specific behavior.

The agent is praising me excessively

Can’t verify
This is a subjective characterization of interaction that cannot be verified without evidence of the actual messages.
ChatGPTDid not respond

Check your own claim

Paste any statement, headline, or AI answer. Independent assessments check it and compare sources.

Share this result