AI, decision-making, and the future of jobs
Future Work/Life is a weekly newsletter that casts a positive eye to the future. I bring you interesting stories and articles, analyse industry trends and offer tips on designing a better work/life. If you enjoy reading it, please SUBSCRIBE HERE, and share it!
When was the last time you made a bad decision? Contrast that with one of the best decisions you've ever made - can you pinpoint what led to these differences in judgement? Did you rush to the poor decision and carefully consider the one you got right? Or was it just luck?
According to Daniel Kahnemann, Cass Sunstein and Olivier Sibony, chances are that 'noisy systems' do at least as much damage as bias in this process, contributing towards unwanted variations in your judgements (evaluations) and decisions (choices).
In healthcare, for example, not only is there overwhelming evidence that physicians, when presented with the same evidence, agree on diagnoses only around two-thirds of the time, one study showed that when individuals' reviewed previous decisions, they disagreed with themselves on up to nine out ten occasions.
The promise of technology and artificial intelligence, in particular, is to help mitigate the influence of human factors like tiredness and mood, on our judgement. However, all too often in business, the conversation becomes about how clever the technology is, rather than the positive outcome of using it.
I've had countless conversations with organisations over the past year about technology and data, at the end of which we inevitably come to the mutual conclusion that, yes, while both of these are important to business growth, they aren't the critical factor.
The desire for companies to shout about their use of artificial intelligence (AI) is a fantastic example of this need to attach a fancy-sounding acronym to the beginning of a product description. You don't need to look very hard to find the newest technology presented as a differentiator, often appended with 'powered-by' - e.g. 'Powered by AI/ML/VR/AR/FARTS'.
All of which misses the point about what genuinely makes a difference to a company's fortunes. We've attempted to reframe this debate by pointing out that the critical determinant to success is, invariably, people. Or since everyone seems to love an acronym, an organisation's HI (human intelligence). Get it? Got it? Good.
For the time being, humans remain uniquely placed to apply creativity, critical thinking, and context. Kenneth Cukier, Viktor Mayer-Schönberger and Francis de Véricourt make this point with tongue in cheek in their new book Framers:
"AI may make better decisions than people and steal our jobs, but computers and algorithms cannot frame. AI is brilliant at answering what it is asked; framers pose questions never before voiced. Computers work only in a world that exists; humans live in ones they imagine through framing."
As they allude to, when it comes to the future of work, there's a narrative that 'robots are coming to take our jobs'. However, on the podcast last week, Azeem Azhar, referencing the emergence of e-commerce and, specifically, Amazon, pointed out that the demise of many traditional retailers was less about the technology itself and more about failures in management.
"If you're one of these companies that hadn't, for example, made the switch to multichannel starting in 1995, when it was apparent that e-commerce was going to be a big thing... then you were going to be ill-prepared for the changes that were going to come. I'm not sure that's a robotisation problem rather than a “we were terrible managers-style problem.”"
So, while we should be mindful of Big Tech's disproportionate power, technology in and of itself isn't a threat to jobs. Instead, if harnessed correctly, in support of human judgement and decision-making, it can be the enabler that helps us reimagine work and, in the process, make it fairer by designing more creative ways of distributing value.
Take, Satalia, for example, a company that positions itself as a thought-leader in AI. Their CEO, Daniel Hulme, was recently a guest on Azeem's podcast, Exponential View, discussing their 'bonkers' (Azeem's words, not mine) approach to creating a decentralised organisation. With the aspiration to develop a genuinely meritocratic, equitable company, Satalia uses radical methods to determine employee remuneration and appraise performance.
If you work at Satalia, not only do you choose each and every project you'd like to work on, judgement on your performance and how much you're paid is determined solely by your peers.
"A recipe for disaster", I hear you say.
However, it has already uncovered and overcome the inherent bias that would have previously resulted in unfair employee pay discrepancies. Initially, the company's policy was for each worker to suggest their salary expectations, which the rest of the team would then adjudicate. In one case, a female developer set her salary expectations significantly lower than her market value, which was caught and corrected by the system. Now, salary-setting goes straight out to the crowd.
What's most notable is the way they use technology to augment human decision-making. As Hulme puts it:
"My hypothesis is that if you give people information, they make good decisions. That's the critical goal. That's what we're trying to use machine learning for - to identify who needs to know what.
We will pull in data across Slack, across our Gmail. We'll do surveys as well to help people understand our strategy. We'll then visualise that in a way that helps them use those insights to then make decisions - whether that's decisions about feedback, or decisions about salary, or decisions about who to hire or not. So we've created this knowledge graph, capturing all this data, and from this knowledge graph, we've surfaced these insights to allow humans to make decisions."
[Above image shows how Satalia talk about the benefits of working there]
In theory, the benefits of this approach include more autonomy, faster decision-making, and more accountability - all of which seems fundamentally a good thing. There are, of course, arguments for centralising some functions, which you can read more about in this HBR article, and it's not completely clear cut that humans even make better decisions the more data we have access to, as this excellent article by Marianne Bellotti argues.
I like the Satalia example, though, because it demonstrates the benefits of taking risks to explore how to improve work, which is becoming increasingly decentralised - both geographically and operationally. As a result, new challenges will emerge that demand innovative solutions for allocating resources and determining pay.
One significant consideration, for example, will be how to reappraise value creation by workers as our traditional understanding of employment continues to evolve - we've already seen a rapid rise in gig-style work and shorter-term contracts over the past year, for example, and that trend will only continue.
The question, therefore, is how soon and how radically we should experiment with redesigning the way we work and how to reward people? Significantly, what role can technology play?
In much the same way as those 90's retailers, failing to evolve will be less an example of the technological impact and more a lack of imagination by the people in charge. So, my suggestion is to start today.
Have a good week,
Ollie
Any Other Business:
One more recommendation from Azeem’s podcast, Exponential View. In this week’s show, Ash Fontana, Managing Director of Zetta Venture Partners and author of new book, The AI First Company, explores the risks and rewards of applying AI to business problems and how to make it a competitive advantage.
More on decision-making in this recent article from Okta CEO Todd McKinnon in Fast Company.
What Do People Need to Perform at a High Level? In this HBR article, Zorana Ivcevic, Robin Stern, and Andrew Faas identified common factors helping (or hurting) employee effectiveness, including individual mindsets and skills compared to managing rules about how work is done and showing emotional intelligence at an organisational level.
How many days should we work from home after COVID-19? The Behavioural Insights Team (otherwise known as The Nudge Unit) release some research into how organisations should approach hybrid work.
The Pandemic Revealed How Much We Hate Our Jobs. Now We Have a Chance to Reinvent Work. So says Joanne Lipman, author of That’s What She Said and a former editor-in-chief of USA Today in Time.
Finally, as we’ve been talking technology today, I had to include this incredible image.
And here it is from another angle.
