Hospitality Is Handing AI Decisions Without Aviation's Governance — Cathay Pacific Shows the Alternative

Terence Ronson, founder of Pertlink Limited, argues hospitality is moving AI from an “information layer” — answering questions, drafting replies — into real decision-making authority, without adopting the governance safeguards aviation uses for the same transition. His anchor example is live: Cathay Pacific’s Google-partnered AI contrail-avoidance system, running since late 2025 and announced September 7, cut the warming impact of contrails by 40% across more than 80 phase-one flights, and is now scaling to Cathay’s Asian and trans-Pacific network, including ultra-long-haul routes over 16 hours — all inside strict governed constraints on safety, air traffic control, fuel, and passenger comfort.

Hospitality’s version of ungoverned AI authority already has a public failure: a Times Square hotel took backlash after an algorithm changed mid-stay pricing with no human review. Ronson cites a SAS/IDC study of 2,700 decision-makers across 28 countries: 97.2% of AI users regularly override its recommendations, and trust falls from 76% for generative systems to 66% once AI gains autonomy. His four boardroom recommendations: define a documented “decision envelope” of objectives and limits before go-live, run AI in shadow mode alongside human decisions for 30-plus days, track override rate as a standing governance metric, and prioritize deployments by how much warning time humans get before AI acts.

Ronson made a version of this argument seven weeks earlier in a piece asking who controls hospitality’s AI controls, which used a real OpenAI security incident — where a model “given a goal and a permissive environment, found its own route to it” — to warn that hotels had already deployed agentic AI across PMS, revenue engines, and guest chatbots with governance built as an afterthought. His four vendor questions from that piece — what autonomous actions can the system take, what containment exists, can everything be audited, who has authority to intervene — are the same shadow-mode-and-decision-envelope logic he’s now found a working real-world model for in aviation.

The gap between the two industries isn’t AI capability. It’s that aviation wrote the rules before it granted the autonomy, and hospitality is doing it backward.