The AI opportunity: Why generative AI is now an unstoppable force

Read time: 6 mins | Added date: 30/11/2023

Bluebells in woods

Generative AI has created headlines around the world and prompted a rush of speculation over how it will disrupt everything from drug discovery to payment processing.

Considered the next generation of Artificial Intelligence, generative AI takes data and uses it to create something new, such as text, images, music, computer code and more.

But how can firms make sure generative AI is being used in a way that’s safe and secure, and what are the key challenges that need to be overcome for it to achieve its true potential? With questions like this front of mind, Lloyds Bank took the opportunity to convene a diverse panel of tech leaders at the recent Leeds Digital Festival to explore this transformative technology.

Existing applications

It’s fair to say that the disruptive potential of generative AI has prompted both excitement and, at times, apprehension among SMEs.

Though there is an awareness about the opportunities that AI can offer, as with any emerging technology, there is also a lack of awareness of how it can be effectively integrated into operations. Against this backdrop, Philip White, Managing Director of software development company Audacia Consulting, opened the panel discussion by sharing a practical public-facing example of generative AI in action.

He said: “We developed a WhatsApp travel chatbot with Northern Rail to provide customers with real-time journey information across 2,500 stations. The generative AI solution, powered by Microsoft’s cloud-based language services, provides customers with live train, station and journey information, which is especially useful during service disruptions. Whilst defining and testing software based on language models can be challenging, the recent developments in language models and cloud-services is definitely making the technology more accessible for organisations to provide improved services through automation.”

Jimmy Forrester-Fellowes, Founder of Rocketware, a software developer specialising in dating platforms, has also been applying AI.

He said: “What’s exciting for me is not so much the customer facing stuff, but how I can use AI to supercharge my systems and enable my staff to do their jobs quicker. I’ve used AI image classification in a photo moderation service, which wasn’t actually any cheaper than workers going through the images themselves, but where the AI has sufficient confidence that the image was safe and suitable, it brought down the execution time from 30 minutes to six or seven seconds.”

Ensuring accuracy

The panel was quick to raise the issue of accuracy and trust, a well-known issue with systems like ChatGPT, which are adept at extracting information from the internet and communicating it in a human-like way, but less so at assessing its accuracy.

One solution is to create siloed systems that can only draw on a pool of trusted data.

Zandra Moore, CEO and Co-Founder of fintech business analytics specialist Panintelligence, highlighted the huge quantities of useful data being held by so many companies which represented an untapped opportunity for AI.

She said: “There is so much incredibly useful data that is not being used and so many low risk, high value applications for that data, like predicting a job that's likely to overrun. There’s been an obsession with data analysis being 99 per cent accurate but I think that in cases like that, companies would be quite happy with an 80 per cent probability.”

Looking ahead, there will be decisions to be made over whether businesses choose to create tailored versions of off-the-shelf models, which will be much cheaper, or build their own models, which will be expensive but more secure.

Regulation and risk

The panel agreed it was inevitable that security will become even more important as regulation comes into the sector but warned that any incoming legislation must be appropriate and tailored to different applications, to avoid throttling innovation and performance.

Systems operating in higher-risk areas, such as cancer diagnosis, for example, should be held to a higher standard than other areas, such as music creation.

Camilla de Coverly Veale, Policy Director at the Startup Coalition, which represents tech start-ups and scale-ups, said that unlike any other technology that’s come before, sector regulators are going to have to really work to understand AI.

And Zandra Moore raised another fundamental risk presented by a lack of diversity in the AI sector at this crucial time in its evolution.

It’s certainly true that the majority of people engaging with this technology at the moment are from privileged Western backgrounds, presenting the danger that their existing bias is being baked into the foundation fundamentals of AI models.

“It will be important for users to contribute to the legislation development process and explain how AI is changing their particular industry, how it’s being applied and how they select one particular AI system over another, because regulators will need to understand this in order to regulate it properly.”

Camilla de Coverly Veale – Policy Director at the Startup Coalition

An unstoppable force

Despite these challenges, everyone agreed that the momentum behind AI is now unstoppable and its capability is increasing exponentially.

Lloyds Bank recognises this and, as part of our work to support SMEs with their adoption of AI, we’ve recently launched A Guide to generative AI, in partnership with UK Tech News, KPMG and Womble Bond Dickinson, designed to help firms understand the applications, challenges and risks around the tech.

Jehangir Byramji, Innovation Lead for Emerging Technology at Lloyds Bank, said: “Anyone who’s used any AI models will see that the rate of improvement over the last 18 months has been so impressive. Part of the reason I’m so enthusiastic about generative AI is that we seem to be at an inflexion point where the capability growth here is phenomenal.”

That means the application of AI will only become more pervasive, both in business and in people’s personal lives.

David Eccles, Founder of digital agency Numiko, envisaged a not-too-distant future where everyone has their own personalised AI that knows their needs and preferences, which we can then trust to carry out tasks like booking a holiday.

He predicted that a big change will happen in the next five years where consumers won’t be using shared models like ChatGPT anymore, but will each have their own personal AI model that has learnt all about them: “That will be the big change that gives us trust and confidence in AI.”

Ultimately it seems unlikely that there will be any corner of the business community left untouched by AI.

At Lloyds Bank, we’re already deploying AI for things like spotting fraud and improving our customer experience, and we’re helping firms of all sizes invest in AI, with applications including resolving customer queries, automating sales processes and improving workflows.

While every new technology comes with a side helping of hype, we can be confident that AI is here to stay. It’s now up to firms to embrace the real-life applications for AI.