Artificial Intelligence: Friend or Foe? | by Etienne Tracey

Since its genesis, the media has tended to portray artificial intelligence as a kind of Deus Ex Machina, an all-knowing and all-powerful source – the future of humankind. The reality of AI, in its application, is often a bit more clunky and prone to human error.

I have had exposure to a range of machine learning technologies since starting at Informa Tech including the work by computer science specialists, event series such as the AI Summit, and automation of marketing techniques. There is no singular Artificial Intelligence, rather each instance requires a specially catered and programmed code that is fit for purpose. 

Echoing the previous Head of Recommendation at Spotify, I have found that our systems are “human all the way down”. In order for AI to ‘become intelligent’, through the process of machine learning, it has to be constantly fed hand-picked data, validated, fact-checked, and cultivated by human interaction.

The benefit of AI, therefore, comes not from an innate intelligence but its ability to sift through trillions of bytes of data, incomprehensible to the human mind. At its foundation, this requires deliberate choice and action of the scientist who programmes it and is, therefore, subject to the whims, biases, and prejudices of that person. More often than not, these are implicit, which means the scientist is unaware of the problems within the data.
Dangers of AI arise in unregulated contexts. For non-computer scientists, there are two things we can question and look out for when we hear about AI: Firstly, Article 9 of the European Union’s GDPR law protects personal identity factors such as race, gender, and sexuality from being targeted with data. We can ask if this is fully being respected by machine learning and not perpetuating any biases. Secondly, we should make sure that sufficient data is used. If not,  the AI will not be able to develop the ‘full picture’. 

By understanding this process, we can see how AI could potentially become harmful – not because of an untapped growth of knowledge and power but because of a particular ‘train of thought’, i.e. data input that holds implicit biases or holes. However, it is not all bad and there are many ways in which we can use AI advantageously. As part of my Multiverse experience, I have used machine learning in order to streamline the data used for marketing, ensuring we appropriately and efficiently meet our target audiences. In the wider context, AI is being used for threat protection, food distribution, and even health care.

Whilst only a select few become computer engineers, it is crucial to have an awareness of the existence, purpose, and impact of AI in our lives and across our professional spheres. When correctly implemented by taking proper governance into account, AI has the capacity to open exciting new business areas.

The common idea of an all-powerful artificial life form is misleading. The illusion of a looming technocratic dystopia is, from my experience, a misrepresentation. Instead, there are already entrepreneurial pioneers who are envisioning the ways AI can be used for social betterment and increasing access to business.

Etienne Tracey is pursuing a digital marketing apprenticeship at Multiverse alongside his job in Audience Development at Informa Tech and is writing for the Apprentice Lens. Here’s more about him:

Etienne hopes to inspire, challenge, and spark curiosity in other apprentices by sharing musings and ideas developed through work. In his words, “I want to explore the bridge between big ideas and real-world applications, where deep thinking meets measurable impact.”

See our full Artificial Intelligence Guide here.