How Are AI & Technological Advances Re-Shaping the World of Business?

At CogX 2022 in London, panellists from Microsoft, QuantumBlack, AI by Mckinsey, and PwC UK gathered to discuss the current and future impact of AI on business. See what they had to say right here…

Artificial Intelligence (AI) is nothing particularly new. These algorithms have been shaping our society and the way we work for years, but only the biggest and most tech-driven enterprises have been able to benefit from them so far.  

But now we are on the cusp of a new era. We’re moving into a phase where accessibility to AI is booming, and more enterprises are adopting it as a result. So how is AI and technological development shaping business? At CogX 2022 in London, Lila Tretikov (Deputy CTO at Microsoft) and Jacomo Corbo (Co-founder & Partner at QuantumBlack, AI by McKinsey) were welcomed by Maria Axente (Responsible AI Lead at PwC UK) to discuss.  

From Optimising Digital to Optimising the Physical World 

Technological development has long seen businesses make strides in optimising their digital ecosystems. But looking to the future, Jacomo Corbo said yesterday that we’re moving to an age where more attention is being given to AI and allowing us to “move the needle on how we optimise in the physical world” – not just in terms of data and processes, but in terms of real impacts on everything from the climate to the way cities operate.  

He gave the example of the energy-intensive steel industry, where ML can help identify minor changes that could optimise the entire production process and drastically cut emissions in the process. What’s more, the ability to learn from successes in this area and translate those insights into other sectors would help break down the silos that hinder collaborative thinking within businesses and across society at large.  

The impact of harnessing the full potential of AI won’t just affect supply chains; it will have implications for ordinary people, too. Lila Tretikov highlighted the fact we’re now facing a future where we have fewer children and young people than adults within Western society for the very first time. While not an issue initially, this will ultimately impact the total workforce available to employers, meaning business leaders will need to find new ways to increase productivity to make up for the shortfall. AI can help with this.  

For instance, in the next few years and as AI and the need for it becomes ever more prevalent, businesses will be looking to hire more engineers to deliver in key areas. There won’t be enough skilled engineers to go around. While technology is at the heart of the problem here, it’s also the solution. Tretikov argued that tech needs to evolve to help engineers build their code. AI is and will become even more capable at writing code and helping engineers and developers to propel their companies forward. 

Nvidia Merlin is just one example. Corbo said that this framework has helped replace hundreds and thousands of lines of code down to just a dozen lines of API code – thereby empowering both engineers and businesses. 

But the potential isn’t limited to code alone. Microsoft pointed out that the level of tasks we can now ask ML to assist with has grown exponentially. Algorithms can now draw pictures, translate conversations in real time, and much more – we’ve gone from simple binary to something far more advanced. 

Answering the question, “Is AI simple plain old IT?”, Corbo discussed how AI has now moved up the abstraction ladder. We’ve improved access to AI, but with that we’ve increased the problems it can pose. Security becomes a bigger risk as more users are able to adopt such complex technologies. But there’s potential for this problem to deepen because developers themselves don’t always know what’s happening behind the scenes. It’s difficult to understand the complexity of what is currently billions of connections that ML makes – and this will only increase as we move towards trillions of connections. Given that training these AIs is already difficult, the problem could spiral out of control.  

AI is clearly have a great effect on the future of business, with more potential to optimise operations, services and delivery than ever before. But where is the oversight of the technology? 

We’re moving to an age where more attention is being given to AI and allowing us to “move the needle on how we optimise in the physical world”.

Jacomo Corbo
Co-founder & Partner, QuantumBlack


Will We See More Regulations to Shape How AI Impacts Business? 

On the regulatory landscape, the EU’s AI act will be a great leap forward, argues Corbo. It will help to shape how businesses build and use AI which is lacking at the moment.

Likewise, Tretikov believes that regulations are a healthy component of business. The best AI scientists can’t hold entire models in their head alone, but regulations could help to provide some level of standardisation or baseline from which to work. This would in turn make it easier to collaborate across regions and providers. The onus, then, should be on regulators to collaborate before bringing forward new legislation to help streamline how enterprises operate globally.

Corbo further said that with the EU’s act, it’s getting a head start on shaping how the world’s AI regulations are formed. Other jurisdictions will “look to live up to EU standards” so this would embed an advantage for enterprises within the region – something we partly saw with the introduction of GDPR.

In the future, expect to see technology adapt to how powerhouse markets like the EU regulate tech like AI – and businesses to plan ahead and meet these regs full on.

Can Artificial Intelligence Ever Fail? 

Short answer: Yes and no. As accessibility of AI increases, we’re seeing a situation where opportunity increases but so too does the security risk these algorithms pose to the enterprise. But it doesn’t stop there; without careful consideration and training of ML, data scientists could end up creating products that don’t positively impact end users or society at large.  

Tretikov raised how Microsoft puts every new product through a process of asset review to validate new products. But what they found in early iterations of the process was that some users wouldn’t raise their concerns, so some use cases and journeys would go overlooked. For instance, women would be less likely to speak up during the review process – thereby removing a significant body of users from the design process. 

The onus, therefore, has to be on creating a more inclusive environment. Review processes need to be structured to address as many user needs as possible. What’s more, while we may be tech evangelists we must also become ethical champions, bringing a design thinking approach to AI development.  

For some, this may start with education. For others, it will be about building trust into transformation.  

In our next article, we’ll look at these areas and more.