Ai Advice - Episode 3
Each week I use our prediction game to test out the world of Artificial Intelligence in the scientifically dubious but often entertaining experiment of asking Google's Gemini AI to make a prediction. Infused with tons of data can a machine with no actual ability to watch the show do better than humans at predicting it?
As with last week, our exchange suggests this might be an area less optimised in its code! 😆 Asked who would win Race Across the World episode 3 Gemini said:
Predicting the winner of episode 3 of Race Across the World gets even trickier because we have even less information. We know the leg goes from Dawson City to Banff, Canada, but that's about it.However, based on the location and the fact it's the third episode, here are some teams that could potentially do well:
So from the start we're not doing well - the AI has got confused with Episode 3 of series 3 - Dawson to Banff, although it seems far less willing to predict that than you would expect for an episode broadcast last year! I wonder if there's a weather forecasting AI that refuses to predict the weather for last week?
Anyway the AI goes on to detailing which teams it actually likes where more confusion abounds.
Teams with good communication: Negotiating with locals, asking directions, and working together effectively become even more important as the race progresses. A team like Sarah & Jess (Mother-Daughter) who seem to communicate well could be strong contenders.
We again have a fictional team, although they sound quite promising! The closest match in the real line up however would be Sharon and Brydie who seem to work together more effectively then Eugenie and Isabel, although I wouldn't currently consider them as "strong contenders"! The AI continues what it looks for in a team:
Teams that can adapt to different travel methods: This leg might involve a variety of transportation options beyond trains. Teams that can adapt to buses, taxis, or hitching rides efficiently could gain an edge. Here, perhaps Alfie & Owen (Best Friends) with their potential comfort with technology for research could shine.
This time we have a real team suggestion, although their comfort with technology seems as useful as their experience in flying planes in a competition that bans both. And I guess being good with tech (and actually in the competition) is less important as the AI concluded:
Leaning on the limited information available, my best guess for who will win episode 3 of Race Across the World is Sarah & Jess (mother-daughter).
Interestingly when pressed for the series prediction the AI responded with:
While I previously suggested Eugenie & Isabel (mother-daughter) as a contender, let's explore a different team for the overall winner of Race Across the World series 4:Brydie & Sharon (mother-daughter)
Which suggests:
A) it is aware of the real-life mother-daughter teams
B) It's expecting a real change of pace for Brydie and Sharon
C) It can't remember who it previously predicted (it has never previously suggested Eugenie & Isabel
Still, there we have it - the official AI advice seems to be "make it up as you go along"!
Think you can do better? Make your Race Across the World Episode 3 prediction here
