What Velociraptors and Mathematicians can teach us about AI

There is currently a lot of hype surrounding AI. Media scaremongering and misinformation abound, often centering around dystopian futures that feature the end of humanity. In my opinion, the underlying source of all the negativity boils down the following: fear of the unknown.

Fear of the unknown is a fundamental fear that all humans share, to varying degrees. The fear is caused by what a study describes as the "perceived absence of information at any level of consciousness." In simpler terms, it is caused by a lack of understanding of the subject we fear.

Another thing we all share is that none of us know (unless you own a tricked-out DeLorean) is what the future will hold. It is one of life's great unknowns. Now pair that with AI. AI is notoriously difficult to explain in non-technical language, as it is a highly complex technology.

When you combine one of our greatest mysteries with a potentially highly disruptive unknown, it is inevitable to feel fear. Now, pepper this mix with a distrust of government and big business and their shaky track record of ethically sound decisions - and it can conjure nightmarish scenarios. A historical example is the fear of Mutually Assured (nuclear) destruction (MAD). A fear that was so great that it spawned a 45-year-old Cold War and shaped society and popular culture.

When you don't understand the source of your fear, you also don't understand the ways the source of your fear can be controlled. And here is where my mind drifts to a different topic: the sci fi blockbuster of the early 90s, the movie Jurassic Park.

One of these experts is a mathematician, Dr. Ian Malcolm. Dr. Malcolm serves as the ethical voice of reason in the film. In contrast to the scientists marvelling at the breakthroughs and businesspeople marvelling at the financial opportunities, Dr. Malcolm takes a humanistic perspective. He warns of the ethical concerns and risks that the scientists who created the park have unleashed. They figured out how to resurrect extinct species and went straight into exploring ways to exploit this technology for financial gain.

Dr Malcolm famously declares:

This quote could easily be applied to what is happening now with Generative AI.

On one hand, you've got businesses rushing to the ‘next big thing', hoping to ride the hype train to quick profits by shoehorning AI into anything and everything. On the other hand, you have the media and its inflammatory rhetoric. Take for example the Time Magazine cover emblazoned with "The End of HumAnIty”. According to certain sections of the media, one would think that we are all on the verge of imminent death at the hands of anthropomorphised robots like Skynet's robotic army. According to some AI vendors, anything and everything can be automated - and their (often vapourware) product is the best one to invest in.

So, what should we do? Should we halt all AI efforts, as the leaders of Open.AI and Deepmind have recommended? Or should we jump on board the hype-fulled gravy train, like so many others are, and recklessly use AI, warnings be damned?

My advice - and brace yourself, it's not very exciting - is… neither. Like all issues, it's not a black and white problem, and the solution isn't black and white either. Each person and business needs to make a decision for themselves on how and why they will utilise AI.

But how do I make that decision? Well, let's see what Dr. Malcolm has to say about that.

The key to rational and ethical decision making is to explore, understand, and learn before making a decision. Again, not an exciting or innovative recommendation. But it is a method that has been proven time after time. We even have a saying for it - "think before you act".

AI, at its essence, is just another tool at our disposal to interpret data. Interpreting data is not a new pursuit. It's what you and I do every day: using our computers. The sleek ones in our hands, the larger ones on our desks, and the squishy pink ones between our ears.

Humans are wired to interpret data. We are also adept at projecting our minds forward and envisioning the potential future outcomes of decisions that have not yet been made. It's called foresight, and it was an evolutionary advantage that helped humanity get to where we are.

Some of us jump into decisions without thinking deeply, and it's at our peril. Humans have the gift of foresight, but some of us are too distracted by the lure of money to utilise it. Just like those that dreamed up the business idea of Jurassic Park.

All significant decisions benefit from a bit of thought and perspective. Whether it’s buying a house, having a kid, choosing what’s for dinner or using AI in our business. For example, how many people would still decide to take out a million dollar loan if they had taken the time to project their future repayments, augmented with statistically likely interest rate rises?

A final quote from Dr Malcolm, as it's the perfect preface for how we should view AI:

AI is not a real human. AI can be controlled. While AI can imitate people, it is not a person. Our tendency to anthropomorphise AI means we mistakenly project all our ideas and beliefs about how humans think and act onto a system that is fundamentally not a human and does not think like a human.

Consider the saying "If it looks like a duck, swims like a duck, and quacks like a duck, then it probably is a duck." Of course, we know that style of reasoning is not reliable. A robotic toy duck can do all those things too. AI might talk like a human, but that doesn’t mean it is a human.

Personally, I don’t find AI scary. When I think about AI, I think about it as a tool. Data goes into it, data and decisions come out of it. In the tech industry, we have the term “Garbage In, Garbage Out” (GIGO). The term originated around the time of early computing. Basically, it means if you give a system flawed data, you’ll get flawed results.

Because this concept has been around for so long, we are really good at mitigating this problem. We have local laws and international standards that we are legally obligated to apply to our systems. Outside of that, we have reams of best practice advice, principles and expertise at our disposal. These safeguards, such as human-in-the-loop are excellent at mitigating risk and it and other techniques work really well to eliminate the GIGO problem.

At CI&T, we have done extensive research on local and global law and guidance. This knowledge underpins how we work. We have synthesised these safeguards into ethical AI frameworks our clients can use. The frameworks are human centred, easy to understand and easy to use. Our goal is to ensure that our clients are able to explore the great benefits of AI and also feel secure that their approach does no harm.

AI that is contained, controlled, human centred and ethical has great potential to revolutionise our future for the positive. Move past the fear by engaging and understanding the topic. Knowledge is power.

I’ll leave you with a final quote, this time from another sci-fi classic, Dune:

---