Digital Economy Dispatch #129 -- Digital Transformation in the Age of AI
30th April 2023
This current obsession with AI is all a bit odd, isn’t in? Seemingly from nowhere, it is all that people can talk about. You know the world is topsy-turvy when you take a London taxi ride and on hearing you are involved in digital technology, the driver immediately challenges you with:
“So, what’s your view on ChatGPT then, Gov’ner?”
Words fail me. But not the taxi driver, it seems, who takes the next 20 minutes to regale me with stories about why the world will never be the same again because of AI. Interestingly, he is a glass-half-full kind of guy and not a doomsday predictor. But your taxi experience may vary from mine.
There is no doubt that discussions about the impact of AI are everywhere. It is quite overwhelming to be bombarded with a torrent of articles, podcasts, radio shows, and everything else. Yet, beneath this hype is a real and important need: To make sense of the digital advances on our doorstep and find ways to position them in the cut-and-thrust of our daily lives. And it seems this is being recognized very broadly.
The swirling discussions take a wide variety of forms. Much of the earlier dialogue at the end of 2022 took the form of “Oh great, look what I just did with ChatGPT”. This was accompanied by an equally strong “oh no, look what someone just did with ChatGPT”. One side focused on the unlimited potential of such technologies to automate repetitive tasks and enhance creativity, while the other was convinced of the unlimited havoc that this may have for employment, the economy, and society. A clash that was highlighted in the recent call for a widely publicised “AI pause” to flesh out this debate.
If the hype surrounding these AI advances has never been more intense, then the need for a deeper understanding of how they are constructed and operate have never been more urgent. Most importantly, it is essential we go back to basics to be constantly reminded of what lies beneath these systems. Unfortunately, in many cases this fundamental understanding of AI is sadly lacking. While the arguments continue, a great deal more attention on AI is now essential to address two additional important aspects: What is it, and what can it do.
In the first case there are several attempts to describe the operational elements of AI based on Large Language Models (LMSs) and why they have suddenly exploded onto the scene. Such descriptions vary depending on their source, from high level overviews in business magazines and broad news channels to detailed reviews and domain-specific analyses from academic research teams.
In the second case, we are faced with an endless set of examples of where and how LLMs are being applied to provide new tools for practitioners. A particularly active areas is healthcare. Here, we see discussions describing how LLMs may be useful in clinical care and health diagnostics, to query large quantities of health data, in resolving health care legal disputes, as the basis for a medical triage service, to improve patient outcomes, and much more.
Perhaps we can offer additional insights here. Let’s take a moment to review some of the context for AI through the lens of the broader digital transformation activities being undertaken in many organizations. Where do these AI advances fit in these on-going initiatives?
It’s Life Jim, But Not As We Know It
As seen in many digital transformation initiatives, the possibilities for digital technology adoption raised over the past 50 years are now being realized by a convergence of advances in data analysis, access to new digital sources of data, high speed connectivity, and raw computing power. Seen together, they are enabling rapid advances in how digital solutions are designed and delivered. The interesting challenge here is to bring this combination of capabilities together to provide what we might view as “intelligence”.
In one typical scenario, more and more data is gathered from a variety of sensors contained in internet-connected devices. By collecting this data, analysis is possible that explores the data to look for patterns. That is, the creation of algorithms to recognize situations and solve problems by learning from earlier experiences and applying that knowledge in unfamiliar contexts. To achieve that feat, what we’re experiencing today is largely knowledge management techniques that use brute force application of very large computing resources to examine extreme numbers of possibilities and variations. By “training” systems with a lot of data about known situations, it is possible to create a form of Artificial Intelligence that compares the new situation to what has been seen before and come to a set of likely conclusions.
To extract economic and social value from the large amounts of information now being stored in global information hubs, we as individuals, communities, and businesses need to convert this powerful and ever-expanding resource into meaningful input that can help us with everyday decisions rather than confuse and overwhelm our lives. The data must support us in addressing key questions: For example, does greater insight into utility usage via smart metering actually improve home comforts? Will the connected car enable us to reduce congestion in cities and avoid accidents? Can banks’ knowledge of financial markets be used effectively to advise us on our retirement needs? Do wearable health monitors lead to earlier interventions to increase wellness and ensure a longer, more active old age?
Technologies such as AI, ML, and MI hold out the promise of being able to make sense of such large volumes of data by exploiting a combination of techniques to yield entirely new sources of value. They encompass natural language processing, image recognition, algorithms, and other techniques to extract patterns, learn from these by assessing what they mean, and act upon them by connecting information together. This is now possible as it builds upon core sets of relatively cheap hardware capabilities provided in massive centres that support large-scale data management (Data Lakes), with the move to virtualized storage and compute power accessible over the Internet (Cloud Technology), and managed distribution networks for architecting efficient systems that stitch together all the pieces of these complex systems (Interconnectivity).
All of this leaves us with an important question that I am hearing more and more: Where does this explosion of activity in AI fit in the broader on-going digital transformation journeys being undertaken in organizations? In particular, how should organizations view the AI advances occurring over the past few months in relation to their existing digital strategies and plans?
Perhaps the most straightforward answer to this query is that it adds emphasis and priority to many of the key elements of digital transformation programmes already underway in many organizations. Efforts to raise the maturity of an organization in the use and adoption of digital technologies have been supported by changes in ways of working to introduce practices that enable new business models, enhance collaboration, and drive experimentation. The resulting digital strategies have tended to focus in three areas that are essential to digital transformation in the age of AI:
Opening up. Digital transformation activities frequently focus on creating more effective communication within and across the organization to improve decision making and speed up action. Consequently, in many cases the need for greater transparency in planning, forecasting, and decision making has forced a great deal of attention on establishing a reliable base of data and defining robust data management practices for improving the quality of data. This effort is critical to successful AI solutions that rely of the scale, scope, and accuracy of the data on which they operate.
Joining up. The disconnected nature of many organizations is a major barrier to efficient, informed operation. Many digital transformation efforts place a key focus on bringing together multi-disciplinary teams across the organization to work cooperatively to deliver value. With emerging AI adoption, a flexible approach to integration across the organization is critical. This demands more agile ways of working that adapt to a very dynamic operating environment.
Smartening up. By becoming more aware of the need for advanced digital skills, many organizations have increased their capacity to manage large data sets, introduced tools to visualize and analyze data more effectively, and supported managers to be more evidence-based in decision making. These form an important basis for a more disciplined approach to AI adoption. The intelligence created as a result of AI adoption relies on effective and efficient approaches by individuals to manage their training, to query them appropriately, and to understand where and how they can ethically applied.
Understanding the basic capabilities of the digital technologies powering AI is important. It is critical if we are to be responsible users of technology and successful leaders guiding an organization’s strategy and motivating colleagues around us. It is not that we all need to be experts in the AI technologies themselves. Rather, we need to know sufficient to be able to consider their implications for what we do and why we do it. To be effective, we cannot abdicate responsibility for obtaining sufficient insight into their operation to ensure they are used appropriately and effectively.
The Digital Way
As discussions of AI and its adoption explode, it is hard not to feel overwhelmed by tools such as ChatGPT and their impact on how we all live and work. Some view this as a kind of alchemy able to change worthless elements into gold. Others see it more like a magic trick, as much sleight of hand as scientific breakthrough. Whatever your position, we should see the latest AI advances as part of the on-going digital transformation journey already being undertaken by many organizations. We can build on these activities and plans to continue to open up, join up, and smarten up our ways of working.