Data Ethics and the Ethics of Data in Safety
I received an email yesterday about data mining and safety data with promises of making unseen risks visible using artificial intelligence (AI). The marketing is ‘choc-a-block’ with promises of what AI can do and how algorithmic analysis can tell you more than you currently know about risk, using the same data.
But let’s get a few basics out of the way first. No data is neutral, just as language is not neutral. All methods hide a methodology (philosophy and ideology) and regardless of branding, one needs to find out what the methodology is before one ‘believes’ or buy’s the spin.
Without critical or cultural thinking Safety stands as an easy target for the snake oil sale. If you want to make a fortune based on naivety, Safety is the industry to target. Until the WHS curriculum (https://safetyrisk.net/isnt-it-time-we-reformed-the-whs-curriculum/) shifts away from managing paperwork to skilling up people in critical thinking, this stuff will go on exponentially. If anyone is going to fall for the machine that goes bing! (https://youtu.be/NcHdF1eHhgc ), it’s Safety!
Promises are easy, just ask any politician but digging deeper than marketing promises is essential if one is on the look out for dehumanizing approaches to safety. When it comes to the marketing of safety the formula is pretty simple: tell a CEO you can improve injury rates with a new lollypop and it’s bound to sell.
Then in a few years when nothing works get in a new marketing campaign and tell them Mum’s for Safety (https://safetyrisk.net/dumbs-for-safety/) will shift the goal posts, well that doesn’t look so good now does it? (https://safetyrisk.net/right-then-children-sit-up-straight-and-take-some-safety/ )
Amazing how this cycle of dishing out cash for slick marketing (https://safetyrisk.net/dumb-ways-to-die-doesnt-work/) seems to substitute for doing anything genuine about safety. In the end people forget the promises made by politicians and all sorts of issues are raised as blame for why something didn’t work then, a new brand of snake oil is sold. And so the cycle goes.
So when you dig deeper into this method of AI and data mining that was sent to me and ask a few critical questions it turns out the methodology behind the spin is behaviourism, why am I so surprised? I saw similar done recently with loads of promises hiding behaviourism masked as neuroscience (https://safetyrisk.net/safety-and-non-neuroscience/). A real giveaway is when you look at the qualifications and experience of the ones promoting this stuff and it’s often engineers and engineering masquerading as social science.
I remember chatting once to a savvy executive about a slick campaign that was being marketed to him and he said: ‘Rob, this isn’t a safety campaign it’s a printing business’. There were: slick graphic designs, little tick cards, folders, 5 for 5 cards, memory stickers, magnets, drink containers, key rings and all sorts of trinkets promising zero. And the cost for this guarantee of zero $2mill! And the previous CEO bought it, 3 years later with a reduced bank account and no change in safety, the new CEO comes in with a different yo-yo (https://safetyrisk.net/the-yo-yo-delusion-and-conversations-about-risk/).
I was reading recently about a spike in cheating in the grand old sport of chess . Seems like online chess and AI have helped out a lot of cheating during the Covid19 crisis. It’s clear that AI in chess works 2 ways and cheating to win is the attraction. At the heart of the problem we find out that the culture of motivation in online chess is set for cheating and a whole new industry of invigilating has arisen.
AI doesn’t sit neutrally on its own but all its algorithms and judgments mask an underlying methodology, of the programmer or designer. The cheating crisis in online chess masks the psychology of motivation in using AI to fight AI.
In safety, I don’t care whether a checklist is on an ipad or an xpad, search for the underlying methodology (driving philosophy) rather than being conned by the machine that goes ‘bing’ or the flash marketing. Unfortunately, many don’t know how to interrogate such stuff or how to search for ethical-political Discourse, motive, power or vested interests. It is astounding how Safety continues to search everywhere for improvements in places where there are none! So, if you get offered a golden turd make sure you scratch below the surface ask a few basic questions.
In SPoR in the module of Social Politics (https://cllr.com.au/product/the-social-politics-of-risk-unit-14/) we offer a simple tool to help this process (see Tool Critical Political Questions below).
It’s not rocket science but if a few more asked such basic questions of the slick marketing stuff about in safety we might actually get back to the basics of safety that is: helping people tackle risk.