AI is much more human than you'd think

5 minute read

As our final module concluded last week, I reflected on the past year and my personal journey on the Oxford Executive Diploma in Artificial Intelligence for Business.

The main learning which came to mind was that it was inherently human. That thought was surprising to me, and I leaned in with curiosity to uncover why.

Indeed, as a Global Head of Research in the Google Creative Works team, I had joined the programme to learn more about the business of data, the algorithms and models, and what makes successful cross-functional teams at the intersection of business and data. Above all, I wanted to share my experience with like-minded professionals and draw inspiration from my peers working in other industries to further innovate with AI in my day job.

Perhaps I felt surprised because we very often focus on the technology, it’s very impressive advances in the past decade and the rapid pace of evolution. For context, when we embarked on our journey in February 2022, text-to-image diffusion models, although developed in the mid-2010s, were at their humble beginnings. Stable Diffusion and MidJourney had not been released and Dall-E was reserved to those enrolled in the beta. Fast forward to the Summer of 2022 and these tools have become a real digital playground - especially for those, who like me, have creative ideas but no executional skills to bring them to life. A future with text-to-video models accessible to the mainstream is now closer than ever. It’s truly fascinating, especially for my industry. My mind has been running a thousand miles per hour thinking about the potential applications.

During the programme, we spent a great deal of time discussing these technologies, the advances, the benefits and most of all the ethics and the responsibility behind those models (see Google’s view on their Imagen tool for more).

Another reason why I was surprised is because we often pit artificial and human intelligence against each other. You can read articles entitled 'AI vs human intelligence', 'Man vs the machine', 'Will AI replace human intelligence?' among other opposing views.

Google Trends: Artificial Intelligence vs. Augmented Intelligence

A quick Google Trends comparison enables us to understand that indeed people are more interested in the technology.

Without further ado, here are the five themes which most resonated with me during the programme:

1. Augmented Intelligence

I believe it is not a question of Artificial Intelligence vs Human Intelligence but Artificial Intelligence AND Human Intelligence.

It’s not an either / or but rather it is about identifying how both intelligences can play to their strengths and how Artificial Intelligence can enhance Human Intelligence. For instance, AI can help automate tedious tasks or analyses, letting humans focus on the most important part of the process which is deriving insights and shaping decision-making.

As one of our esteemed professors wrote in his book: ‘Don’t compete with AI at the things it’s good at, complement AI with the things you’re good at’ - Martin Schmalz and Uri Bram, The Business of Big Data.

I am very energised about unlocking precious time and more strategic areas for our human intelligence to shine further and be used at its highest potential.

2. AI is just a tool – albeit a disruptive innovation

I am always amazed by the technology and how powerful it is. It is one of the most disruptive technologies and innovations of today and will revolutionise industries and reinvent the way we do business. I do not mean to undermine it but as with any technology, it is just that: a tool to help us achieve the outcome we want in the most effective and efficient way.

It represents a bundle of opportunities and capabilities. However, we need to understand what to use it for, when, how and most of all why. Some people like to refer to Machine Learning as ‘glorified statistics’ in the age of big data. It may not always be the answer. Sometimes a simple linear regression does the job.

It’s important that as humans, we focus on the question or business challenge we’re seeking to answer and define if AI can help us get there. Let’s not ‘do AI’ for the sake of it! It’s a common business trope these days.

3. Human-centric design

Businesses create value by delivering human-centric goods or services - AI is a tool in service of a human-centric product offering. It’s always been about human experiences and AI can help enhance this and drive value for our customers and therefore business.

4. Human-in-the-loop

Having a human-in-the-loop (HITL) is best practice. From setting the strategy and the business challenge that the model will help answer, to deriving insights and taking decisions based on its findings, humans are critical at every stage, from inception to implementation.

People are critical to success when embracing an AI opportunity. The organisational challenge is often a human one vs a technology one. It’s about getting management buy-in, breaking down the silos, influencing at all levels in the organisation and infusing a disruptive mindset.

It’s very much about business strategy and leadership in the age of AI. Same humans, different technological challenges.

5. Ethics and bias regulations

Humans are critical to regulate the use and potential misuse of AI, and to protect society against bad actors while ensuring most benefit from its advances in all areas (health, climate, social…to name a few).

When thinking about potential biases, it’s important to recognise the AI learns from datasets which contain biases that are inherently human. It would not be biased per se, but, unfortunately, it replicates and perpetrates existing human bias. Think about the fact that the media, and advertising for example, is not as diverse as, or representative of, the overall population. AI will learn from our human bias and project it a thousand-fold, leading us to take suboptimal, adverse decisions should we not question the input and consequently its output.

Furthermore, I deeply valued the human connection and bond we formed as a group as we came together in the world-renowned and fascinating Oxford University. I loved learning from our world-class faculty, guest speakers and fellow classmates who bring a wealth and breadth of experience and insights - all 34 of us bring about 512 years of cumulative experience across 19 industry sectors (from tech to the military).

This is why I found that this experience and AI was truly human.