Artificial Intelligence is being hailed for its potential to transform many aspects of our lives – in particular, it provides a big opportunity for us to do new things and do them differently - But we need to understand the implications of this at global, national and local levels, in a systematic way.
The UK government has rightly recognised the importance of AI in developing a modern industrial strategy and is seeking to keep the UK at the forefront of the application of these new technologies. For example, the Department for Science, Innovation and Technology (DSIT) recently announced the creation of The Sovereign AI Unit to build and apply AI capabilities to drive economic growth and national security. But a key question is whether the UK has the right strategies in place to realise this promise.
While many in industry and government are getting excited about AI as a key to problems across the country and society, there is a lack of integrated analytical thinking on the matter. The challenge in doing this systematically is that much of this debate is already dominated by two distinct mindsets, which largely reflect the bi-polar view of technology-driven innovation:
- the technology mindset which aims to promote the revolutionary potential of AI, emphasizing how recent developments, particularly in large language models (LLMs) have transformed the landscape beyond the foundations in this field laid over the last 50 years: this view is promoted aggressively by the major technology companies who would like to dominate this space, and the associated eco-system of new players.
- the investor mindset, which sees this as the ‘next big thing’, which could transform the fortunes of investors and investment funds by delivering financial returns outside the typical expected envelope: this has been aided and abetted by governments and policymakers who recognise the potential to move the needle on the growth agenda.
The big problem, however, is that the revolutionary nature of this potential transformation needs to be understood in the context of a multi-polar challenge, not just technology and funding. AI has the potential to fundamentally transform market structures and competition (hence the concerns about ‘data scraping’), the regulatory landscape (driving the cross-national battles over the nature and scope of AI) and the ways new products and services are created, manufactured and deployed (leading to governments placing greater focus on IP ownership and the creation of new data centres).
To tackle this systematically, we need an analytical framework which allows us to explore and understand the different dimensions of Artificial Intelligence. Without a coherent integrated framework, we risk addressing this challenge like a group of blind men trying to identify an elephant; only capable of understanding one aspect at a time.
Our unique analytical framework, which builds on the 8-layer model for contingent (or conditional) technology deployment (developed as part of the Triple Chasm Framework), attempts to address these challenges by explicitly articulating the different components of AI technologies as pervasive ‘horizontal’ (or digital) technologies, focused ‘vertical’ technologies and hybrid ‘cross-over’ technologies at the intersection between the two. This provides the basis for a multi-polar analysis of the promise of AI.
Unpacking Pervasive Technologies
AI Technologies represent the latest incarnation of pervasive digital technologies with the potential to affect a very wide range of products and services and market spaces in different and sometimes complex ways: this raises complex and un-precedented challenges for governments, businesses and individuals in terms of the economic, social and political impact; probably like the dislocations created by previous information-related technologies, such as the internet and mobile telephony.
We also expect similar challenges to manifest themselves in the future as we start to address how quantum technologies are commercialised, which reinforces the need for a structured approach to understanding the contingent nature of technology exploitation and deployment. Figure 1 below summarises how AI technologies can be deconstructed and characterized vs the generic 8-layer model. In particular, it allows us to unpack the different technology components, to understand the interaction between the components, and to understand how they can be integrated to enable new functionality.
Figure 1: AI seen through a Pervasive Technology Lens
Understanding Focussed Technologies
While the 8-layer contingent technology deployment model works very well for pervasive technologies, it does not account for different domain contexts where they might be deployed. While pervasive digital technologies are often ubiquitous, focused technologies present a different challenge for two reasons:
- the wide range of technologies across different domains including biology, chemistry, physics, materials science and engineering.
- the possibilities for these focused technologies to be integrated to enable new functionality (e.g. identifying new biomarkers in medicine).
We tackle this challenge by defining market-space-centric value chains, as described in the Triple Chasm Model and building on the conventional understanding of Porter’s value chains. The key is to define how value is added, which then provides the basis for identifying the relevant focused technologies: these technologies may cover an extremely wide range of science and engineering, including physics, chemistry and biology.
In particular, we need to understand how these technologies impact:
- The different value-adding elements in any specific market space value chain.
- Linkages between these value elements and how they can be integrated to deliver different outcomes.
- Understanding the changing nature of these interventions covering both linear and non-linear behaviour with feedback loops.
Hybrid Technologies: cross-over between Pervasive and Focused Technologies
It turns out that the greatest value creation from artificial intelligence will actually be found in the creation of hybrid technologies where pervasive technologies interact with focused technologies to create new capabilities. This means that the key to understanding the opportunities and constraints of AI depends critically on understanding the detailed value chains of specific market spaces, ranging from drones to electrical vehicles and new diagnostics and therapies in healthcare.
At the general level, the potential impact of these hybrid technologies can be understood based on the generic value chains associated with the creation, development, testing, deployment and optimisation of new products and services, which is why there is much confusion and debate about the design, deployment, regulation, and business models of AI-enabled products and services.
Market-space centric value chains can also display varying degrees of complexity and potential impact, which is why any rigorous analysis needs to be based on the relevant detailed value chain, with an understanding of the critical role played by data and meta-data. In particular, this is critical if regulators are going to understand and tackle the challenges in different market spaces: for example, the argument advanced by the big technology companies for why regulators should not constrain acquisition and use of data to create better models, could end up having catastrophic consequences for those who create new forms of content as a key part of their business rationale.
Unbundling the Impact of AI in Specific Sectors (Lifesciences and Healthcare)
The value of these hybrid technologies can be best illustrated through a more nuanced assessment of AI by looking at how the pervasive technologies associated with AI can be mapped against a complex market-space-centric value chain to generate focused new technologies which can form the basis for new products and services. This approach can be applied to every market sector from the Media and Entertainment Industry, with new possibilities for how content is generated, packaged, distributed and delivered, to the Defence Sector, where AI is dramatically changing how conflicts are conducted.
For this article we have chosen to focus on the Lifescience & Healthcare market-space-value chain, as shown in Figure 2. This sector is one of the more complex market spaces to illustrate but has become a core focus of the governments industrial strategy with an ambitious goal for the UK to become the 3rd most important life science economy in the world.
This emerging value chain contains 11 main value-adding components, starting with advances in genomics, including genotype-phenotype mapping, leading to new diagnostics and therapies delivered to patients in a variety of delivery settings.
Figure 2: The Emerging Lifesciences & Healthcare Market-space-value chain
We can then map the pervasive technology analysis for AI technologies discussed above versus this value chain to identify potential hybrid technologies where new products and services can be created. Figure 3 illustrates this mapping and highlights different types of hybrid products and services. The examples shown are not intended to be exhaustive, but they illustrate several different generic types of products and services enabled by hybrid technologies which combine pervasive AI technologies with focused technologies associated with different parts of the value chain.
These new hybrid technologies occur all across the value chain. Earlier stage developments can be found in the new insights that can be gained from RNA-focused research & Bio-marker discovery and their incorporation into personalised medicine. Later stage developments through the analysis of large datasets, are enabling new developments in the management of clinical trials, individual patients and entire hospitals.
Figure 3: Hybrid AI Technologies in Lifesciences & Healthcare
Broader Implications
This structured approach to understanding how AI technologies are commercialised is critical for policymakers as they continue to develop industrial priorities. They must, however, take care to not be blinded by the black box hype presented by those with narrow interests. Understanding all of the components of AI technologies as a pervasive technology is vital for understanding many of the broad implications for the industry and how the UK can best position itself. To best exploit many of the opportunities with significant potential for growth will also require incorporating domain expertise of the new and existing focused technologies within key market sectors so promising hybrid technologies can be identified and supported effectively.
The country finds itself in an excellent position to achieve this, but must deploy its expertise appropriately, in a structured and integrated manner that is not blinded by the hype from tech visionaries or investors with no incentive to consider the big picture. The Industrial Strategy Sector (IS-8) Plans, released so far have provided solid footing for companies in each sector but, unfortunately, the implementation of AI has been limited to the Digital and Technologies Sector Plan and broad initiatives like the skills agenda. As we progress into the implementation stage it is vital that policy makers focused on each sector in the IS-8 identify the high potential intersections where hybrid technologies may be found and supported.