Fluency: Why Data Diversity has an important role to play in driving inclusion
Fluency’s view is that data diversity will drive better inclusion because it weeds out the unconscious bias that we have as human beings
Fluency M&C Saatchi recently offered thoughts to Barclays touching on the topic of data diversity.
Why Data Diversity?
Diversity has only truly come into focus in recent years, triggered by unacceptable global headline news and unsavoury experiences being shared with shocked audiences in many parts of the world.
In a world where social media is part of daily lives, many of us are exposed to richness and diversity on a scale never known before, offering endless benefit and opportunity. There are also less positive responses to diversity which are well-publicised – including the risks of turning our backs, living in bubbles, choosing to reject realities that do not suit.
At Fluency, we talk about data diversity. We are, of course, aware and careful not to cheapen the meaning and importance of the diversity discussion. Diversity, though, touches on data in important ways, and can help deliver important lesions to those who work with data at any level.
Our view is that data diversity will drive better inclusion because it helps weed out the unconscious biases that we have as human beings
Fluency, in Building Brands Better – Barclays
First, much has now been written about whether some of the newly launched AIs are really capable of understanding diversity – and the dangers that this can bring to decision-makers who are relying on them. We are in agreement that this needs close consideration. Every use of AI should be considerate not only of what we are learning – however powerful – but also of what might be being missed, and why. At Fluency we think of insight very much as a science, something that should be endlessly challenged, questioned, deconstructed.
Second, we must also be ready to accept that some new AIs offer scale, depth and speed of discovery that could only be dreamed of just a few years ago. We are still regularly surprised and delighted by the new opportunities we uncover here at Fluency – and highly conscious that, even if 10 people are working on a project, our collective lived experiences are not even the tiniest fraction of the scale of data we are working with, let alone hope to understand without the help of machines. Fluency ‘lean in’ to this new capability as a tool wisely. We believe that we should expect AIs to teach us as much as is possible about a large spectrum of thinking that exists out in the world. Through rational, scaled and data-diverse AI assisted approaches, we can expect to learn new things about ourselves and each other and start to uncover more of what previously lay impenetrable to human capacity alone.
Third, many of society’s best efforts to achieve diversity depend on bringing together difference and variety wherever possible. This is exactly what we do at Fluency, with when it comes to the data we expect and hope to learn so much from. We do not think decision-makers should be relying on any one type of data to understand the world around them.
We construct solutions that are inherently diverse in input; whether combining behavioural data, claimed data, written data, audio or image data, our objective is always to weed out the unconscious biases that narrow or solo data sets or AIs can bring.
When we hope to learn new things about humans, consumers, markets – it is through these combinations that we can best hope to reveal our world’s many truths and lived experiences.
This is what we mean by data diversity.