Byte Size Digital episode 2: How to harness Artificial Intelligence

ANDDigital podcast: How to harness AI

Craig Mundy (Leeds Building Society) and Adam Hall (AND Digital) discusses how to harness Artificial Intelligence and how to get the best out of generative AI tools in episode 2.

The second Byte Sized Digital podcast by Leeds Building Society & AND Digital is out now. It scrutinises Artificial Intelligence (AI) asking questions like “Are we going to lose our jobs to AI?” and “Is AI going to take over?”

Hopefully if that doesn’t scare you then you can take away some tips on how to get the best out of generative AI tools and hear some opinions from our subject matter experts.

A brief background to Artificial Intelligence

When the first computer chatbot, ELIZA, was created in the 1950s, it required hardware costing $3.5 million and ran on a mere 15 bits of memory. Fast forward to today, and for just $75, you can purchase a Raspberry Pi capable of running an open-source Large Language Model that’s billions of times more advanced than ELIZA. It’s safe to saythings have moved on a lot.

We’ve now achieved consumer-grade AI capabilities, igniting a media frenzy about its implications. Companies are racing to harness these machines, seeking an intelligent system that provides a competitive edge. The adoption of these systems is skyrocketing. For instance, OpenAI’s ChatGPT became the first digital product to amass 1 million users in just 5 days – a feat only rivalled by the unsolicited U2 album distributed to iPhone users in 2014.

Will it take away my job?

As more individuals begin using these tools, a growing concern has emerged: “Will AI take away my job?”.

Consider technical revolutions from the past. The printing press made the jobs of hundreds of scribes redundant. But the fact that the printing press existed not only gave rise to jobs in construction and maintenance of the machines themselves, they also created a whole industry as it lowered the barrier to entry of recreating books. We saw the rise in the role of “author” and “publisher” and “journalist” – all of which weren’t an option before.

The invention of the internet may have been destructive to some industries (like music, retail, publishing, transport) but it also gave rise to new industries (like social media, online retail, digital music and communications).

“People are worried about these things….I think in the short to medium term, it will actually increase the number of jobs.”

On the podcast, Khaled Mahmood Leeds Building Society Business Analyst highlights the emerging role of Prompt Engineering that we’re seeing more frequently. Prompt Engineers are adept at instructing AI models to execute time-consuming tasks, acting as translators between human intent and machine execution. Mastering this role demands an intimate understanding of AI systems and their inherent flaws.

How do I become a Prompt Engineer?

To get good at using these tools you need to understand its limitations.

An example we all need to be aware of is the susceptibility of Large Language Models to “hallucinate” – which is where the AI will present information that looks real but is actually just a sequence of words which don’t have any real world context.

The beauty of prompt engineering is that it demands quintessentially human qualities. It’s the ability to describe tasks, to converse with people, to discriminate between viable solutions.

These are the things that make us better prompt engineers, and they are the things that make us more human. What truly excites me is the potential to offload numerous administrative tasks to AI. We will get to apply it to work in ways that give us higher abstractions on the work we do, which can mean we are ultimately more productive.

Just as Excel Engineer was once a sought-after role, the Prompt Engineer job title might be transient, eventually becoming a standard expectation for the digitally adept worker.

What does this mean for the future?

It’s hard to predict where we ultimately end up, but it’s safe to assume that tools like Generative AI being available to society will have a profound impact on the way we collaborate and communicate. We’re seeing thousands of companies spotting gaps in niche implementations of the technology – so whilst ChatGPT is a general tool, more people are trying to apply it to very specific domains to create services for micro communities.

We are also likely to see a change to the way we interact with companies. For decades, industries revolved around SEO (Search Engine Optimization). This is the act of crafting content so that our website will appear higher on search engines than our competitors. Essentially we have spent 20 years writing content to appease Bots (Search Engines).

Now we have the ability to create content using bots. A future where bots are writing content for bots isn’t going to be great for users, so I predict a website renaissance where we will witness new ways of interacting with the services we use, and with the advent of Generative AI, improvements in voice recognition and the adoption of smart speakers, I would say that conversational interfaces are likely to be key in this coming transition.

The podcast is hosted by Craig Mundy (Senior Delivery Manager) who is joined by Steve Love (Principal Development Engineer), Khalid Mahmood (Business Analyst) and Adam Hall (Principal Technologist).

To hear more views discussed by our seasoned AI SMEs and enthusiasts watch the episode here. You can also listen here!

All events

All sponsors