Upstream is a joint research initiative from Fieldcraft Studios and Braidr, and the founding project of Braidr’s AI for Good programme. It maps the gap between what UK wild swimmers, paddlers, anglers and surfers care about and the antimicrobial resistance (AMR) sharing the water with them, and publishes the findings free for anyone designing awareness work. I led the Braidr side and owned the analytical pipeline end to end.
Why “stop swimming” doesn’t work
Antibiotic-resistant bacteria have turned up at every UK river site sampled, and University of Exeter research found surfers carrying resistant E. coli at three times the rate of non-surfers. Most people know about sewage in rivers. Almost nobody connects it to drug-resistant infection. And the people most exposed are the most attached to the water, so anything that reads as “stop swimming” gets binned.
The research asks what each community already cares about, whom it trusts and how to connect AMR to those concerns. The format matters too: the same people will read a paragraph on Reddit and two sentences on TikTok.
How it works
I built the analysis pipeline around a system prompt grounded in published AMR science, risk-communication research and UK regulatory context, with each claim cited inline. Every analysis call uses that prompt. The model propagates those citations into its outputs, so the analysis traces to the Exeter study and Environment Agency surveillance rather than to whatever the model half-remembers. Every behavioural signal also gets a strength rating from strong to isolated, so one striking post can’t masquerade as a trend.
The public site covers ten communities across seven dimensions, from opportunity sizing to message hooks close to finished copy, with a detailed page for each community.
The wild-swimming pilot on TikTok set the method. Across 482 UK posts from 169 creators, AMR came up zero times, and the dominant safety heuristic was whether the water looked clear, the most dangerous and most correctable misconception in the data. Those findings became five messaging approaches, each explaining how to connect audience concerns to AMR, who could credibly deliver the message and what risks to consider.
Research is ongoing, with collection still underway on some platforms. The project is part of the Braidr work recognised at the 2026 DataComms Awards. Colleagues built the collection pipeline and visual design; Fieldcraft brought the brief and AMR expertise, and turns the matrix into campaign creative.