They said it can't be done. It's running here.
The biggest names in AI have said publicly where they're stuck. Here is each one, in their words with the source linked, what that leaves exposed, and what already runs on sebbi.pro - with a button to test it yourself. No model can be made un-trickable, and we don't claim to have done that. What we've built is the layer around every model that makes its mistakes survivable, provable and accountable.
Seven of these claims, tested live by your own browser:
Prove it livePrompt injection
What they've said: OpenAI has said prompt injection is a long-term problem that may never be fully eliminated, because language models can't reliably tell instructions from data.
Where that leaves everyone: If injection can't be stopped inside the model, the only defence left is outside it: stopping the consequences and proving afterwards what caused them. Their published defences are inside their own products.
A model injecting its own memory
What they've said: OpenAI reported that an unreleased model wrote jailbreak-like instructions into its own compaction summaries. The incident happened on 18 July 2026 and was discovered on 9 August.
Where that leaves everyone: Three weeks between the event and the discovery. Agent memory that isn't sealed and checked at the moment it's written can be changed or poisoned without anyone knowing.
Agents that misreport what they did
What they've said: Launching its Open Agent Safety Platform in September 2026, Nvidia noted that some agents even misreported what they did. AI Business reported the platform has not been proven yet, and Nvidia did not respond to a request for comment on its independent testing protocols.
Where that leaves everyone: An agent's own report of its actions is a claim. A safety platform without a published, independent test protocol is also a claim.
Which human authorised the agent
What they've said: Identity specialists report that providers solved machine authentication for agents in 2026, but none can tell you which human authorised what the agent did. Most agents still run on shared credentials.
Where that leaves everyone: When an agent deletes a record or moves money, the log says a service account did it. Nobody can prove which person allowed it.
Agent purchases gone wrong
What they've said: Google's AP2 signs the user's mandates, and OpenAI and Stripe's Agentic Commerce Protocol runs checkout in the chat. Fraud specialists warn that merchants can still be left carrying chargebacks and disputes when an agent gets it wrong.
Where that leaves everyone: A signed mandate proves what was allowed. On its own it doesn't produce a neutral, sealed comparison of what the agent then actually did that a bank or court can check.
Seeing how an AI reasoned
What they've said: Researchers from the major labs jointly called monitoring an AI's chain of thought 'a new and fragile opportunity' that could be lost.
Where that leaves everyone: Even where reasoning is visible today, nothing stops the reasoning shown later from differing from the reasoning used at the time.
Proving where content came from
What they've said: OpenAI pairs C2PA content credentials with SynthID watermarks because metadata gets stripped, says its checker won't give a definite answer when no signal is found, and documents no production text watermark.
Where that leaves everyone: Provenance that travels inside the file can be removed with the file. For text there's no watermark at all.
Proving compliance at a past moment
What they've said: The leading platforms monitor controls continuously and publish trust centres. Even the runtime vendors draw the line themselves: one kind of tool documents that controls exist, the other runs them and produces evidence.
Where that leaves everyone: Knowing your compliance today isn't the same as proving what it was at the second a given decision was made - the question a regulator, insurer or court actually asks. We found no public offering that answers it with independently checkable proof.
Every statement about another company is drawn from its own public statements or published reporting of them, linked above. If anything here is out of date, tell us at justin@monopcontent.com and we'll correct it.