AI and human, an inevitable new division of labor

by Didier Mamma — VP Data Value Creation

Clearly AI is changing our everyday life as the steam power or electricity revolution have done before. So the question is: How to make sure that AI and Humans enrich each other and not competing against each other?

Here after I have synthesised a framework that we are experimenting at Decathlon. The framework has been inspired by Kai Fu Lee and tuned to be actionable.

MAKE A BRIDGE CROSSING THE CHASM FROM BOTH WORLD

So the good news is that Human beings have overcome such revolutions. However the current revolution is more complex for 2 reasons

• Our planet cannot sustain anymore an uncontrolled exponential development
• AI is challenging what make Human so different than the rest of species i.e. his brain.

As a result AI has to “assist” humans to fix the awfully equation complexity by maintaining a rapide business development, environnement protection and last by not least people happiness. The consequence is that we cannot avoid to rethink the division of labor with a new eye.

ASK TO YOURSELVES THE RIGHT QUESTIONS

It is a very basic principle in science i.e. make sure that you characterise carefully the problem you want to solve otherwise the probability to bring good answer to the wrong question is high.

DISRUPT THE WAY OF THINKING

The challenge is to find the right balance between too simplistic when it comes copping such topic or put too much complicatedness to be actionnable.

Simplistic / Reductionist approach: not recommended

Having such “2 dimension brutal force” approach the temptation would be to automate via AI 60% of tasks i.e. repetitive — Routine and Optimisation. such decision would be a terrible mistake triggering huge resistance of change from employees as well as from customers.

Systemic approach: recommended

Let’s “humanised” classification adding 2 adding variables: “the need for empathy / trust vs no need for empathy and trust”. The outcome is very different revealing 4 Categories.

• Routine job: Tasks largely automatisable requiring limited or no inputs from Humans e.g. Accounting, Fraud analyst, Compliance…
• Augmented Analytics Intelligence: Tasks where AI feeds human Thinking thank to complex computational data exploration e.g. fundamental research. Product engineers, Engineering…
• Performance Optimisation: Tasks where AI Analytics brings choice of several next best actions, simulation, what if scenarios e.g. Revenue manager, Category Manager, Supply Chain manager, Radiologist, surgery…
• Augmented Emotional Intelligence: Tasks where AI augments empathy capacity in order to anticipate better people desire e.g. Game designer, Marketing Director, Police officer….

TAKE AWAYS

1. Ask to yourselves the good questions.
2. Find tune the equilibrium between simplistic and complicatedness.
3. Keep in mind that a Human does understand the goal of what IA does. AI does not have goal neither understanding of what and why it does what it does. Which imply that IA should never ever be a blackbox otherwise avoid it.
4. Link with the point #3 never ever neglect the continuous effort and time for the change management (unlocking resistance). I include reskiling and up-skilling programs.