In this second opinion piece, I’ll be looking at the steps that standards makers should be considering as their customers increasingly adopt and rely on AI-powered systems, and in particular generative AI (Gen AI) tools. And this is set against a broader challenge of remaining visible in an ever-increasing sea of content publishers and digital disruptors.
Let’s be clear – Gen AI is a digital disruptor, and like most digital innovations this creates specific opportunities and threats for standards publishers. While many standards publishers are in the process of evolving their digital strategies, some may still rely on traditional workflows, such as PDF formats. However, this presents a valuable chance for transformation, including the use of further content licensing options. Most key customers are early adopters of technology, and many are digital-first organizations. The Gen AI disruption is widespread and pervasive – and being felt from healthcare to automotive to services to defense and beyond. In short, there are few industries that are not in some way being impacted by Gen AI’s ability to create new content, and in some cases insight, from existing data.
The first step is to understand a little about why GenAI is so attractive to your customers. The capabilities of GenAI AI are significant. In content generation, for example, GenAI can produce summaries, translations, and content visualizations, streamlining the creative process and reducing time and resource investment. It can improve natural language processing, so it enhances communication technologies, from chatbots to sophisticated translation services, facilitating smoother, more natural interactions between machines and humans. Businesses are finding it easier and more efficient than ever to create designs and even products that were previously unfeasible due to time or resource constraints. Put simply, the perception is that Gen-AI powered tools can help save time, money, and free up employees to move further up the value chain. There are challenges, too, and it is here where standards makers probably have the most to gain.
A key opportunity with GenAI is that it is highlighting more than ever the need for standards – in both digital and human arenas. Key standards here include ISO 42001 Artificial Intelligence Management System and IEEE Ethically Aligned Design, both creating significant interest in the market. On the digital side, GenAI is nothing without data (most importantly, copyrighted content) – it relies on algorithms that in turn rely on schemas, ontologies, interfaces, and impacts many other technical areas such as cybersecurity, management systems, and most obviously software development; on the human side it accentuates the need for verification and validation of data, and increasingly, ethics and ‘human-in-the-loop’ decision-making. Who’d have thought that we’d be having to legislate to keep people as those ultimately responsible….
Those standards makers who have already been working on digital infrastructure, and in particular on digitalization of content, will benefit most from GenAI. The combination of metadata and content creates even more value – well-structured and semantic data can power discovery tools like conversational agents. A more commercial approach is to make standards content easier to integrate into learning systems (i.e., language models). In this latter scenario we mean language models deployed as internal knowledge bases, or domain-specific ‘small’ language models – not the big general-purpose beasts that get all the headlines.
Purpose-built, niche, verified, validated. Standards makers are ideally placed to provide this kind of content (or partner with those who can) and maximise their IP commercially. Whether it’s ‘SMART’ standards or content-as-a-service, these GenAI technologies create the potential to completely reimagine the user experience for standards, with virtual assistants providing access to standardized knowledge in critical use cases, from discovery to assessment. The challenge will be to do this in a way that creates meaningful value for users, and not simply creating a more expensive sales funnel. This will mean deliberate curation and possibly revision or extension of the content that trains these agents – e.g., a careful appraisal of when to build this in-house, when to partner, and what technologies to adopt. There are an increasing number of credible solutions, and the cost of training a language model to meet your specific needs continues to fall – although the skills required to develop and manage these workflows are still relatively scarce. You may find yourself competing with your own customers for resources – perhaps co-creation is the route in 2024?
Proactive engagement in the AI revolution is not just about mitigating risks; it’s also about seizing these opportunities to innovate, to look again at your business model and re-evaluate where the value in your organization is – hint, ‘it’s your content and data – all of it’. Customers are also looking to standards to ensure the quality and reliability of their AI applications and also to boost public trust in AI technologies. In short, being AI-ready and proactive is not just a precautionary measure for SDOs; it’s a strategic imperative to harness the full potential of AI for both standards makers’ own businesses through effective licensing, while ensuring its safe, ethical, and effective deployment across the globe.