‘This is the stupidest these machines will ever be.’

These are the words OpenAI CEO Sam Altman remarked when being interviewed by Bill Gates at the beginning of last year; honestly, it was a mic-drop moment.

Think about all the AI you’ve used, read about, heard being mentioned by colleagues and friends, and seen demos on LinkedIn. 

Now take a moment to read that quote again and let the sentiment sink in.

It’s fair to say that most of us will have had a bit of an AI mind-blown experience over the past few years. For many, it was the “Oh wow, AI really is magic” moments - like when ChatGPT-3 went mainstream and suddenly everyone was an AI expert.

Because, here’s the thing, Altman wasn’t kidding, the pace of development has accelerated further. There have been a tonne of new market entrants and model upgrades, and China’s started to show its hand. Multi-modal (using speech, images, text, etc) became a thing. Increased reasoning took off. Agentic AI is everywhere. And the quality of outputs continues to leap forward.

The Parameters the models are trained on have skyrocketed. So much so that the likes of Google, Microsoft and Amazon are placing big bets on small nuclear reactors to keep pace. And for many of us, interacting with AI, whether consciously or subconsciously, has become ingrained in our daily routines.

So, all good progress, aside from the carbon footprint, yes? Well, yes and no.

The devil’s in the detail

As the tech’s improves, and we use it more frequently, we gain more value. We automate more of the mundane and find more time for more human-y, intellectually challenging aspects of our businesses. Critical thinking, innovation, transformation and change, and softer skills allow us to do all this with empathy. 

At the same time, we often feel compelled (by FOMO) to explore the latest AI releases that promise even better results as they take another step closer to human-level intelligence. However, there’s a danger we forget to question its outputs and the inputs that inform the results.

And this is critical.

We’ve already suffered from racist robots, GPT hallucinations presented as evidence in courtrooms, businesses tanking because AI crunched the numbers wrong and biased recruitment nightmares. And, when we consider the adoption of AI in education, healthcare, policing and government, the stakes only get higher.

Regulatory progress is being made, with the EU AI Act trying to keep pace; however, as with all things tech, the regs lag behind the pace of development. At the same time, there’s been an alarming trend of those in big tech quietly axing their ethics teams.

As such, self-regulation is becoming increasingly important. 

Time to take responsibility

It’s no longer a case as to whether a business will adopt AI, it’s now very much when and how its use will be formalised. 

Your opportunity is to create some differentiation in the race for all things AI. Take a step back and ensure that your business is powered by a responsible AI approach that builds a deeper level of trust in your organisation.

Formalising AI use or putting the governance in place doesn’t have to be an arduous task but, it is critical for two reasons:

1. Without these guardrails, your team can, unwittingly, be putting the business at risk. They may be uploading sensitive information, breaching GDPR, violating copyright and potentially perpetuating bias and societal inequalities.

2. Not having the right governance means you’re not using AI responsibly. If you haven’t formalised your approach to ethics, security, safety and privacy, your team may wonder why and question if you care. A lack of clarity about the red lines that won’t be crossed and what AI adoption could mean for people’s roles will see them vote with their feet.

Putting the governance in place

There are three considerations that flex according to business size:

1. Governance Framework - Defining the core principles - so everyone knows how AI gets used, tested, and monitored. Ethics, safety and explainability should be baked in, not bolted on.

2. Governance Committee - Establishing a stakeholder group to review and govern AI adoption, usage and approval. Ideally, this would involve at least one person from outside your business to provide an objective view with different learned experiences. In larger businesses, consider involving appropriate NEDs/Trustees and academics, where possible.

3. Governance Policy - Formal and published documentation including the governance framework, foundational principles, processes, management, reporting, working groups and escalation. Once in play, think about how you communicate this to your supply chain partners and what aspects you need them to sign up to.

   
   

Keep it simple, work out the best approach for your business, get the right foundations in place, define some guardrails, and go from there.

Evangelists, evaders and everything in between

Look around your business and you’ll find different mindsets, outlooks and worldviews. That’s what keeps things interesting, right? But it also applies to AI. Not everyone is an eager early adopter. Different people will have different comfort zones, and some will actively avoid, either due to fear, a lack of understanding or ‘not being a techie’.

With this in mind, it’s vital to take everyone on the journey with you. Communicate the business’s approach to AI, hold forums, educate people and dramatise some of the amazing use cases that have already had a positive impact on society. Think, new drug discovery, adaptive learning, wildlife conservation, food production and even its use in the hands of artists and filmmakers.

Sometimes it’s useful to reframe AI as Augmented Intelligence. A less threatening alternative that implies a more bidirectional relationship with the tech. Or as Fei-Fei Li, known as the Godmother of AI, puts it:

"Artificial intelligence is not a substitute for human intelligence; it is a tool to amplify human creativity and ingenuity." 

Don’t know your AGI from your DNN?

If the whole “doing AI responsibly” thing feels daunting, let’s chat. 

We’ve got the strategy, governance know-how, and proven AI advisory chops to help you navigate this emerging world. 

Because if we’re going to play with the most intelligent tools ever, we might as well do it smartly.