Skip to main content

Part human, part machine: Finding the balance between people and AI

Girl touching hands with robot

With advanced technology, are we losing our purpose in understanding humans?

Big data – and what to do with it – has been a conundrum for researchers for years. We feel like we have failed if we don’t figure out how to capture the value of immense amount of consumer data available to us. In response to this challenge, we have seen a proliferation of research tools emergent in the world of SaaS. These technologies promise to harness the extensive available data putting the power of AI and machine learning in researchers’ hands. For many researchers, this is an exciting prospect.

The risk in our race to adopt ‘ResTech’ is that we potentially lose sight of important information relevant to our endeavour of better understanding people. While it’s important to embrace technology and push the boundaries to find new and better ways of doing research, the significance of the human element in research cannot be understated.

Advanced technology can be used to augment human methods in innovative and symbiotic ways, but not replace them. We still need humans to get human insight. 

So, in assessing and defining the role and purpose of ResTech tools, there are some important factors for researchers to consider.

Available data is incomplete

A key consideration when assessing the role of new research tools is that the available data used in isolation fails to show the full picture. This applies to any pre-existing or available data such as social media, customer feedback data or customer transaction data. These data sets can have a purpose and be revealing, but they fall short if our purpose is to try and better understand humans to develop more impactful consumer strategy and make better decisions.

A human is necessary to meaningfully interpret the data sets available to us.

The AI tools available to theme social media data, for example, are hugely powerful and can save time for researchers. Technology can theme social media data effectively; it just can’t provide meaning to those themes. Tools can’t replace the researcher and the strategic thinking and analysis that needs to be overlaid to the data for a full and complete interpretation.

AI can’t explain human behaviour

Importantly, these tools don’t account for human behaviour. They don’t reveal the ‘why’ behind what people do. AI can mimic human behaviour in some basic cases but equating that with understanding human behaviour is like saying a Polaroid camera helps us understand human vision. Behavioural science is a necessary complement to AI.

That is because we know that humans are social and massively irrational. People don’t do what they say or say what they do. Behavioural science is developing at an exciting pace. The field is revealing more about what is at the core of human behaviour, going beyond opinions or rationalisations – which aren’t connected to the way people really act – to inform strategy.

You can’t separate people from the context in which they are operating

The ultimate context that influences people is culture. This is what they are swimming in every day, the common discourses they are exposed to and that affect what they think, how they act and the decisions they make.

Machine learning and AI tools can help to broadly curate and theme what is happening in culture, providing a shortcut which can be carried forward into other forms of research, where we can explore these themes and how they are manifesting in the context relevant to the research question. Advanced technology tools can be a powerful springboard into strategic thinking and developing platforms to take into research and explore in more depth the context and nuance around those themes with real people.

For example, social media needs to be looked at in a cultural context. Social media is a rationalized and deliberate sharing of experiences so reveals little in the way of behaviour or motivations. It’s not complete, it’s often not true, and it’s hugely influenced by the context in which the posts are shared.

While it’s valuable to analyse what is being shared on social media, what is not being shared can often be where the real insights into people are found. This deeper level of understanding can only be revealed by interacting with and observing people to understand the relationship between context and behaviour that is actually occurring. You need both parts to see the full picture.

The way forward for AI and machine learning

There are exciting things going on in tech which we should embrace and explore, test and improve, but we need to understand their limitations and make sure we’re not sucked in to the point where we miss the point of what we’re doing. When we layer human analysis, by a human, onto rich data from advanced tools, we start to unlock truly powerful insights.

Going forward, when approaching advanced technology, we need to maintain this balance; part human and part machine.

This article first appeared in the August-October 2022 edition - Human Insights - of The Research Society’s publication Research News


Mark Hobart
Mark Hobart
Managing Partner, TRA Melbourne

back to top

Discover more content

Stay in touch!

Sign up to receive our latest thinking straight to your inbox each month.