Artificial intelligence powers a lot of what makes this app interesting. Here is exactly what gets generated by AI, what data is used to generate it, and what AI cannot do in this app.
The following features are AI-generated. They are produced on demand using large language models and are not written by a human.
A satirical camp newspaper written about your group. Uses guest names, quest results, meal ratings, scores, and tips submitted by group members.
An AI-written radio script voiced by Gord Mackenzie of CKBJ 980 AM. Covers the same ground as the Bugle. Audio is generated by ElevenLabs.
Recipes generated from a meal title, serving count, dietary restrictions, and the kitchen equipment listed at your venue. Not sourced from any specific recipe database.
18-track playlists generated from a vibe description and session type. Songs are suggested by the model — they are not pulled from a licensed database and may include songs that don't match your taste.
Award citations written for each winner based on their actual trip activity — quest completions, scores, Bugle coverage, meal ratings.
When a guest submits a side quest, the AI evaluates the submission against the quest description and rules. It makes a pass/fail decision. There is no human review.
Nearby attractions, area history, and local context generated from your trip location. Based on the model's knowledge of the area — not a live directory lookup.
A preview edition of the Bugle published before the trip starts, introducing the group and building anticipation. Uses registered guest names and trip details.
To be specific about what is not AI-generated:
When AI features run, a prompt is constructed and sent to an external AI provider's API. The prompt includes structured data from your trip. Here is what is included:
Beyond names, the data included varies by feature. The Bugle prompt includes quest completions and outcomes, meal names and ratings, leaderboard standings, submitted Bugle tips, activity schedule, trip dates, location area name, and group vibe/content rating setting. It does not include chat messages, photos, PINs, email addresses, or the property street address.
Recipes include kitchen equipment list, serving count, dietary restrictions, and the meal title. No guest names are included in recipe prompts.
Quest judging includes the quest description and the guest's submission text. The guest's name may be included for context.
AI-generated content in this app — particularly the Morning Bugle and local intel — will sometimes be wrong, strange, or confidently incorrect about facts. This is a known characteristic of large language models, not a bug specific to this app.
The Morning Bugle is satire. It exaggerates, misremembers, and invents. Gord Mackenzie does not know what actually happened. He has a strong opinion about it anyway. This is the point.
Local intel and area history are based on the model's training data. They may be outdated, incomplete, or incorrect about specific local details. Always verify anything safety-relevant (road conditions, fire bans, wildlife advisories) through official sources.
AI setlist suggestions are based on the model's knowledge of music, which has a training cutoff date. Songs suggested may not exist, may be misattributed, or may simply not match what you asked for. The setlist is a starting point, not a guarantee.
The organizer sets a content rating before the trip starts: PG, mixed adult, adults-only, or party. This setting is passed directly to the AI as an instruction that adjusts tone, humour, and language in generated content.
At adult or party settings, the AI is instructed to produce content that is deliberately irreverent and may include crude humour, profanity, and adult themes. This is an explicit product feature — it requires the organizer to actively set the rating. The default is PG.
We do not moderate the output of adult-rated AI content beyond the instruction set. If you set an adult or party rating, expect adult content.