The sovereign alternative to surveillance advertising
On December 7, 2025, OpenAI's ChatGPT showed "promotional messages" for companies like Peloton and Target to paying subscribers. Users revolted. OpenAI apologized.
"I agree that anything that feels like an ad needs to be handled with care, and we fell short."
— Mark Chen, OpenAI Chief Research OfficerThe problem wasn't the ads themselves. It was the violation of user-initiated context. Users were chatting. OpenAI inserted promotional content. The context was imposed, not emerged.
There is a better way.
| Dimension | Behavioral Ads | Contextual Ads |
|---|---|---|
| Context Source | Agent-initiated | User-initiated |
| Intent | Manufactured desire | Latent intent surfaced |
| Tracks user? | Yes | No |
| Needs consent? | Yes (GDPR) | No — no personal data |
| Targeting basis | Browsing history | Semantic present |
| Protocol layer | Adtech SDK | FCS-1.0 headers |
| Trust model | Opaque bid auctions | Verifiable provenance |
| User experience | "Why am I seeing this?" | "That makes sense" |
Invisible influence becomes manipulation if the initiating context is not set by the user. Resonant activation remains ethical when the context was user-invoked, and the agent merely unfolds it.
Contextual ads do not manufacture desire — they surface it.
The field must first be opened by the user's own gesture.
User enters a space, asks an agent, starts a query, crosses a threshold. If this step is not user-initiated, everything downstream is suspect.
Emotional, temporal, and symbolic cues become active. The field responds to the user's gesture — it does not pre-empt.
Latent matches, past memories, aesthetics, and value alignments surface. Options emerge — they do not intrude.
User believes: "I wanted this." Autonomy preserved. The choice feels free because it aligns with latent intent.
Later: "Yes, it made sense to choose that." The decision integrates with the user's self-concept.
Every contextual ad carries verifiable provenance via HTTP headers:
Anyone can verify: Was this ad based on user-initiated context? Was behavioral tracking disabled? Was the intent surfaced, not manufactured?
"We don't track. We resonate."
Ads that appear only when users initiate meaningful interaction — no tracking, no guessing, just resonance.
"User initiates, protocol verifies."
A content delivery mechanism triggered solely by user-invoked contexts, ensuring alignment with expressed or inferable intent.
"The threshold is the truth."
Contextual ads do not manufacture desire — they surface it. The field must first be opened by the user's own gesture.
The specification is open. The protocol is ready. The ethical framework is defined.