Research outputs
Publications
Formal papers on AXI. Each published paper has its own landing page with the full text and bibliographic metadata. Work in progress is listed separately on the roadmap and is never shown here until a public version with a persistent identifier exists.
Artificial eXperience Intelligence (AXI): An Engineering Discipline for the Human Experience of AI Systems
Published preprint Version 2.0
| Author | Yeluri S. D. S. Sri Vardhan, Sripto Corporation Private Limited, Andhra Pradesh, India |
|---|---|
| Type | Preprint (not yet peer reviewed) |
| Server | Research Square |
| Version | 2.0, posted 28 September 2026 (supersedes 1.0, May 2026) |
| DOI | 10.21203/rs.3.rs-9717445/v2 |
| Licence | CC BY 4.0 |
| Full text | HTML with figures and appendices · PDF · Markdown |
| Cite | BibTeX · RIS |
Abstract
Artificial intelligence has crossed a threshold in how it presents to people. In pre-registered three-party Turing tests, a persona-prompted language model was judged human more often than the human it was compared with; listeners can no longer reliably tell cloned AI voices from real ones; and such systems now drive cars and act for absent users. As capability rises, the decisive question is no longer whether AI can communicate like a person, but whether it communicates with people comfortably, usefully, and honestly. We propose Artificial eXperience Intelligence (AXI), an engineering discipline whose central construct, the AXI Bridge, is a runtime control plane between any AI capability and human cognitive bandwidth, governed by the rule of human-grade naturalness with machine-grade honesty. AXI specifies a five-layer architecture with API contracts, nine design principles, seven telemetry-anchored metrics, and a multiplicative score whose consent and integrity gates cannot be averaged away; an autonomy profile adds bystander-legibility and safety-case gates for vehicles and robots. In a retrospective analysis of eight documented incidents (2018–2025), AXI's hard requirements would have constrained the harmful action in three and its metrics target the failure signature in three; two lie outside its scope. Operationally, a live tutoring platform recorded about 3,000 sessions in one week, with roughly 90% of that week's users returning within it, and developers informally observed a positive reception of a near-human avatar in a 150-participant pilot; these observations are descriptive, not evidence of effectiveness. We specify four validation studies and an inter-rater reliability check for pre-registration.
Citation
Yeluri S. D. S. Sri Vardhan (2026). Artificial eXperience Intelligence (AXI): An Engineering Discipline for the Human Experience of AI Systems (Version 2.0) [Preprint]. Research Square. https://doi.org/10.21203/rs.3.rs-9717445/v2
In progress
Two follow-on outputs are being drafted (a Bridge technical specification and a guardrails conformance benchmark) and two depend on data that has not yet been collected. See the research roadmap for their questions and status. None has a public version yet.