I was asked to comment on a few GPT-related questions by Node4’s PR agency. I haven’t seen where those comments are featured yet, but I decided to string them together into my own post…

How has ChatGPT prompted innovation over the past year, transforming industries for the better?

ChatGPT’s main impact in 2023 has been on driving a conversation around generative AI, and demonstrating the potential that this technology has to make truly impactful change. It’s certainly stoked a lot of interest in a lot of sectors and is clearly right at the top of the hype curve right now.

It’s sparked conversations around driving innovation and transforming a variety of industries in remarkable ways, including:

• Revolutionising Customer Experiences: AI-powered chatbots, bolstered by sophisticated language models like ChatGPT, can engage with customers in natural language, offering real-time assistance, answering queries, and even providing personalized recommendations. This level of interaction not only improves customer satisfaction but also opens new avenues for businesses to understand their customers on a deeper level.
• Enhancing Decision-Making and Strategic Planning: By harnessing the power of AI, leaders can make informed decisions that are driven by data rather than intuition alone. This has impacted decision-making processes and strategic planning across industries.
• Transforming the Economy: ChatGPT, with its human-like writing abilities, holds the promise of automating all sorts of tasks that were previously thought to be solely in the realm of human creativity and reasoning, from writing to creating graphics to summarizing and analysing data. This has left economists unsure how jobs and overall productivity might be affected [as we will discuss in a moment].
• Developer Tools: OpenAI and Microsoft have unveiled a series of artificial intelligence tool updates, including the ability for developers to create custom versions of ChatGPT. These innovations are designed to make it easier for developers to incorporate AI into their projects, whether they’re building chatbots, integrating AI into existing systems, or creating entirely new applications.

These advancements signal a new direction in tackling complex challenges. The impact of ChatGPT on workers and the economy as a whole is far-reaching and continues to evolve. 2023 was just the start – and the announcements from companies like Microsoft on how it will use ChatGPT and other AI technologies at the heart of its strategy show that we are still only getting started on an exciting journey.

Elon Musk has recently claimed that AI will end all jobs, is this actually a reality or is it scaremongering?

Mr Musk was one of original backers of OpenAI, the company that created ChatGPT; however he resigned from their board in 2018. At the time that was over conflicts of interest with Tesla’s AI work and Musk now has another AI startup called xAI. He’s well-known for his controversial opinions, and his comment about AI ending all jobs is aimed to fuel controversy.

Computing has been a disruptor throughout the last few generations. Generative AI is the latest iteration of that cycle.

Will we see jobs replaced by AI? Almost certainly! But new jobs will be created in their place.

To give an example, when I entered the workplace in the 1990s, we didn’t have social media and all the job roles that are built around it today; PR involved phoning journalists and taking clippings from newspapers to show where coverage had been gained for clients; and advertising was limited to print, TV, radio, and billboards. That world has changed immensely in the last 30 years and that’s just one sector.

For another example, look at retail where a huge number of transactions are now online and even those in-store may be self-service. Or the rise in logistics opportunities (warehousing, transportation) as a result of all our online commerce and the fuel for ever more variety in our purchases.

The jobs that my sons are taking on as they enter the workplace are vastly different to the ones I had. And the jobs their children take a generation later will be different again.

Just as robots have become commonplace on production lines and impacted “blue collar” roles, AI is the productivity enhancement that will impact “white collar” jobs.

So will AI end work? Maybe one day, but not for a while yet!

When and if it does, way out in the future, we will need new social constructs – perhaps a Universal Basic Income – to replace wages and salaries from employment, but right now that’s a distant dream.

How far is there still to go to overcome the ethical nightmares surrounding the technology e.g. data privacy, algorithm bias?

There’s a lot of work that’s been done to overcome ethical issues with artificial intelligence but there’s still a lot to do.

The major source of bias is the training data used by models. We’re looking at AI in the context of ChatGPT and my understanding is that ChatGPT was trained, in part, on information from the world-wide web. That data varies tremendously in quality and scope.

As it becomes easier for organisations to generate their own generative AI models, tuned with their own data, we can expect to see improved quality in the outcomes. Whether that improved quality translates into ethical responses depends on the decisions made by the humans that decide how to use the AI results. OpenAI and its partners will be keen to demonstrate how they are improving the models that they create and reducing bias, but this is a broader social issue, and we can’t simply rely on technology.

ChatGPT has developed so much in just a year, what does the next year look like for its capabilities? How will the workplace look different this time next year, thanks to ChatGPT?

We’ve been using forms of AI in our work for years – for example predictive text, design suggestions, grammar correction – but the generative AI that ChatGPT has made us all familiar with over the last year is a huge step forwards.

Microsoft is a significant investor in OpenAI (the creators of ChatGPT) and the licensing of the OpenAI models is fundamental to Microsoft’s strategy. So much so that, when Sam Altman and Greg Brockman (OpenAI’s CEO and CTO respectively) abruptly left the company, Microsoft moved quickly to offer them the opportunity to set up an advanced AI Research function inside Microsoft. Even though Altman and Brockman were soon reinstated at OpenAI, it shows how important this investment is to Microsoft.

In mid-November 2023, the main theme of the Microsoft Ignite conference was AI, including new computing capabilities, the inclusion of a set of Copilots in Microsoft products – and the ability for Microsoft’s customers to create their own Copilots. Indeed, in his keynote, Microsoft CEO Satya Nadella repeatedly referred to the “age of the Copilot”.

These Copilots are assistive technologies – agents to help us all be more productive in our work. The Copilots use ChatGPT and other models to generate text and visual content.

Even today, I regularly use ChatGPT to help me with a first draft of a document, or to help me write an Excel formula. The results are far from perfect, but the output is “good enough” for a first draft that I can then edit. That’s the inspiration to get me going, or the push to see something on the page, which I can then take to the next level.

This is what’s going to influence our workplaces over the next year or so. Whereas now we’re talking about the potential of Copilots, next year we’ll be using them for real.

What about 5 or 10 years time?

5 or 10 years is a long time in technology. And the pace of change seems to be increasing (or maybe that’s just a sign of my age).

I asked Microsoft Copilot (which uses ChatGPT) what the big innovations of 2013 were and it said: Google Glass (now confined to the history books); Oculus Rift (now part of Meta’s plans for augmented reality) and Bitcoin (another controversial technology whose fans say it will be world-changing and critics say has no place in society). For 2018 it was duelling neural networks (AI); babel-fish earbuds (AI) and zero-carbon natural gas.

The fact that none of these are commonplace today tells me that predicting the future is hard!

If pushed, and talking specifically about innovations where AI plays a part, I’d say that we’ll be looking at a world where:

• Our interactions with technology will have changed – we will have more (and better) spoken interaction with our devices (think Apple Siri or Amazon Alexa, but much improved). Intelligent assistants will be everywhere. And the results they produce will be integrated with Augmented Reality (AR) to create new computing experiences that improve productivity in the workplace – be that an industrial setting or an office.
• The computing services that are provided by companies like Node4 but also by the hyperscalers (Microsoft, Amazon, et al.) will have evolved. AI will change the demands of our clients and that will be reflected in the ways that we provide compute, storage and connectivity services. We’ll need new processors (moving beyond CPUs, GPUs, TPUs, to the next AI-driven approach), new approaches to power and cooling, and new approaches to data connectivity (hollow-core fibre, satellite communications, etc.).
• People will still struggle to come to grips with new computing paradigms. We’ll need to invest more and more effort into helping people make good use of the technologies that are available to them. Crucially, we’ll also need to help current and future generations develop critical thinking skills to consider whether the content they see is real or computer-generated.

Anything else to add?

Well, a lot has been made of ChatGPT and its use in an educational context. One question is “should it be banned?”

In the past, calculators (and even slide rules – though they were before my time!) were banned before they became accepted tools, even in the exam hall. No-one would say today “you can’t use Google to search for information”. That’s the viewpoint we need to get to with generative AI. ChatGPT is just a tool and, at some point, instead of banning these tools, educational establishments (and broader society) will issue guidelines for their acceptable use.

Postscript

Bonus points for working out which part of this was original content by yours truly, and which had some AI assistance…

Featured image: generated via Microsoft Copilot, powered by DALL-E 3.