"Those who don't see themselves may feel cycling is not for them": Images of cyclists lack diversity and focus on cycling as leisure rather than transport, researchers suggest

There's another issue with the distribution (or not) of images of different demographics cycling that is more subtle and possibly worse than just not representing minorities. AI training sets will typically include large sets of internet images (they're free, so companies will use them) and this means that any bias will end up getting encoded in the weightings. As AIs get more and more hype, police forces are likely to get suckered in and start using them for such tasks as traffic policing and the inherent bias will then show up by the AIs discriminating against non-white/non-male cyclists and flagging them for stop and search and the like. This of course will then feed back into official figures and show that minorities on bikes are more likely to be involved in crime (because they've been targetted by the AIs) and so we end up creating a fictitious state of affairs by unthinking use of biased images.

And there's the issue of trained AIs being used to create "new" images which will likely be posted on the internet and further enforce the bias. AI will eat itself.

There is another worrying message in these findings:

"Though no official data for the cycling industry exists at the moment, the report also noted that, anecdotally, Black, Asian, or people from ethnic minorities are underrepresented in the sector, as are individuals from low socio-economic backgrounds or with disabilities. There is also a lack of LGBTQ+ people joining or leading the UK bike industry, the report said."

I suspect that this is partially because members of minorities (that have often suffered from forms of hidden racism or abuse) are trying to blend in, avoiding things that are not mainstream (and may raise criticism from the bullies), trying to be "normal" - as a result of society putting pressure on them to "assimilate"... (This is only my impression, and I only have anecdotal evidence, but if this is true, then the even bigger concern should be about how members of the mentioned minorities are treated than about them cycling less!)

This is another piece of research trying to find a reason for poor uptake in cycling in the UK which finds predominantly white male healthy (skinny if you're doing research from a driving seat) being heterosexual middle class and drinking only cappuccino with almond milk. 

It's nothing to do with psychotic drivers, police indifference, motornormativity, bias in CPS, judges and the motor insurance industry. Nor has it anything to do with a capped 0.001% of the transport budget for well-designed cycle infrastructure since 1890.

Given the quality of this 'research' I'd not brag about, instead actually talk to people who cycle instead of doing an AI farce.

I find the term "family cycling" biased, as it already implies a leisure activity. (Most parents work, kids attend school - so the only time for cycling together is, normally, leisure time.) - However, I can see why they did not search for "cyclist" instead - as this shows people on roadbikes. (Which again may seem fair enough, as these images will be more "symbolic" for a cyclist, more easy to recognize and more distinct than a "housewife locking up her bicycle at the super-market to do the shopping".

It's also worth noting that the level of diversity in the google results may well depend on ones own search history (but this is hard to test unless you compare to friends with a very different background - culture, place, language, ...)

Trying to assess how badly the images found by "family cycling" lack diversity compared to other search results, I searched for "driver". My results showed a "good level" of diversity - but there was another bias, which was quite shocking to me: All these drivers seem to smile - which certainly does not seem to reflect reality - being trapped in a box, in traffic, getting insufficient excercise or fresh air. (It's a bit like car-adverts that always show an impossibly empty road. They are more likely to show a bicycle than even one other car!)

Oh, the grifters have arrived. Don't let them in, keep the door locked shut. They get only paid to divide people and won't stop finding things to divide when their grift depends on it. It's just a bike so just ride it.

Not quite sure what the news is here?

• The demographic mix of UK cyclists doesn't reflect UK society.  Reasons will be complex and may vary for different kinds of cyclists / cycling (e.g. I'd bet there's a positive skew away from "UK origin" in the ethnicity of folks working for the food delivery companies).  There may be less skew in "mass cycling" places e.g. NL rather than the UK.   That might give us a clue as to a good way to address things (not just changing pictures).  There although "locals" cycle more than "immigrants" the latter certainly cycle.  Oh, and uniquely more trips are made by women than men!
• There is bias in images of cyclists - also not news.
• There is bias in images of stuff in the world.  Really not news.

Do the images reflect the realities of who's cycling?  And if not why not?

Being a vanilla dude "I don't know what I don't know" but I can't help feeling there are a couple of more fundamental barriers to cycling... "being able to visualise it" is important, but question:

Just how much would changing the pictures (even if it was to reflect reality better - see the first question) really make a difference to numbers cycling in the UK?

People in the UK give all kinds of reasons why they don't cycle.  Skepticism needed - asking about things people previously haven't considered may get you speculation and confabulation... However the top reason given is likely close to the reality: all kinds of people just don't cycle because it's not "safe" (pleasant) or convenient.  Then: people don't cycle because very few others (of any gender, age or ethnicity) cycle.  Cycling for transport in particular is not perceived as a social activity (and in the UK this is partly by design).  Finally many people have a car, right there, so the question doesn't really arise for many...

82% of the people living in the UK are white, so most of the people who cycle will be white.

Isn't this more a question of Google or any other search engines algorithm not selecting images that reflect.the diversity of people cycling rather than there being a massively heavy preponderance of skinny white heterosexual families that cycle.

I haven't read the research in detail (obvs) but "family cycling" is a pretty blunt search. It's not surprising that it returns e.g. lots of images of rural cycling away from traffic - because most of the images appear on leisure and tourism sites. If you were Visit Hampshire, would you use an image of the urban commute through Southampton, or a leisure ride in the New Forest? (There may be a geographical flaw in this argument, but you get the point)

EDIT: I have now read it, it's here. I accept the general premise that cycling images are relatively undiverse; that people who don't see themselves in images of cycling may be in some way dissuaded; and therefore that it is desirable actively to increase diversity in images of cycling. But a search for "family cycling" does not, I would suggest, provide a representative sample of cycling images people might be exposed to. If, for example, you were trying to find about about local cycle routes, you're unlikely to come across and be dissuaded by these images because you wouldn't search for "family cycling" because it's untargeted.             

They searched for "family" cycling. Is the cause of their upset that they then found representations of kids having both parents in their lives, out together cycling, as a family? And, moreover, enjoying the health benefits. What were they hoping to find?