Is Google Biased In Favour of Cars?
Later this summer, we’ll be making our annual trip to Penzance to see my mum.
It’s a journey I’ve made every year for 35 years, so I know the options pretty well, but let’s see what Google has to say:
This is interesting as, according to Google, the quickest option for getting to my mum’s flat in Penzance from my house in Walthamstow is to drive.
Now, I know that’s not true but it’s interesting that Google says it is as it is symptomatic of a subtle little bias embedded into the way Google (and every journey planner) works.
People eat and sleep
The Google algorithm is realistic about walking time, interchange time and connection times with public transport, but not the equivalents with a car. Google assumes that I will drive non-stop for 5hr 38 minutes. In reality, I’ll almost certainly need to stop for a meal and it would be deeply irresponsible not to stop for a break.
It also assumes that I can just pull up outside my mum’s flat; the reality is that I need to find somewhere to park.
In practice, this journey should assume a 45 minutes break (as per regulations for hauliers and coach drivers) and at least 10 minutes to park. That extra 55 minutes makes a big difference, as public transport becomes the fastest option.
Now, you could argue that I might swap driving with my wife part way and so not need to take a break. I might have access to a private parking space when I get there. But, of course, as a public transport user, I might cycle to the station as opposed to walking. I might walk faster through the station than Google assumes.
For both public transport and private transport, it’s reasonable to assume a certain norm of behaviour.
It sounds like a small issue but there are so many examples in which our system is hardwired to subtly prioritise cars.
If someone from Google gets to read this blog, next time you update your algorithm, please make it fair between cars and trains. Or, to put it another way, given cars’ impact on the climate… don’t be evil.
What do you think? What other examples of similar biases can you think of? Tell me on LinkedIn
Do you Tweet? Here’s one ready-made
