- Better cycling infrastructure. Yeah, no shit. We need more paths, more direct paths, and more connected paths. You should be able to go anywhere you want using a route that it at least as direct as the most direct driving route, by bike, without ever sharing a road with cars above 30 km/h, and with a minimal number of road crossings where the cars get priority.
- Use AI to identify where cycling infrastructure needs to go. 🙄 Or you could just ask cyclists. We’ve got no shortage of ideas of places that are severely lacking already. Maybe the AI could be useful once most of the basic network is done, but not today.
- Improve transport modelling to include cycling. Yes! Add in induced demand effects on infrastructure for cycling, public transport, and cars. Use models that understand traffic evaporation when reducing road widths or adding modal filters. Our transport engineers are currently woefully behind the times.
- Politicians need to actually care about cycling. Yeah, no shit.
- Make active transport funding a priority. Yup. Our councillors love to harp on about how they spent X amount on cycling infrastructure, but they never put that in context of how much is spent on roads. But also, let’s make sure that money goes where it’s most useful. Spending billions building one green bridge is great, but is still much less useful than building many kilometres of good separated bikeway for the same price. (The real answer is to do both!)
- Recognise the health benefits of cycling. Yes, but this isn’t really an actionable item. It’s just more reason to do the above items, particularly pointing to 4.
Use AI to identify where cycling infrastructure needs to go. 🙄 Or you could just ask cyclists.
I guess you didn’t finish reading that section:
A big advantage of AI is it can be scaled up. Once trained, AI models can be replicated across many neighbourhoods to identify urban design features that support cycling. It’s even more useful when combined with citizen science and rider experiences, as we plan to do.
No, I just think it’s silly to talk about applying AI to something that just manifestly does not need AI. It’s a dumb buzzword at best, an excuse to spend less money actually building infrastructure because more money is going to AI consultants at worst.
Like I said, if it were about filling in the little cracks once we have a really good overall network, I could maybe get behind it. But right now there’s just zero need for it, because the stuff that’s missing is so obvious and there’s so much of it. At least in Brisbane, the Council could decide to triple its spend on bike infrastructure and still take a decade before the big problems we’ve been calling for action on for years are all exhausted.
an excuse to spend less money actually building infrastructure because more money is going to AI consultants at worst.
How will using AI in a privately funded research project take money away from government funded infrastructure projects?
