Whether organizations are currently throwing themselves wholeheartedly into the AI melee or not, developing and implementing a coherent strategy and policy around the technology is a must given its unremitting advance.
This is the view of Kriti Sharma, Chief Product Officer of Legal Tech at Thomson Reuters and Founder of AI for Good, a social enterprise that develops tech for social good. The idea here is that whether companies are ready or not, AI is coming, not least as it is already being embedded into many software suppliers’ applications. She explains:
All organizations should have an AI strategy and policy. It could be a ‘turn your business around in the highest priority areas’, a ‘go slow to go fast’ approach or anything in between, but it’s important to have something. It’s about educating employees and implementing processes so that people don’t just end up using generic Large Language Models or enter insensitive data into systems. But it’s also about enabling colleagues and teams to discover and learn from the technology. In other words, it’s about laying the foundations.
Doing so makes sense before AI adoption moves into the enterprise mainstream, Sharma believes, which despite all the recent hype around generative AI, it has yet to do. As she points out:
This is a peak of the AI movement, but it’s not the peak and there will be more to come. This is the equivalent of being at the dial-up phase of the internet, so while the technology is great in some areas, it still needs to mature and grow in others. Therefore, for tech leaders, it’s currently about prioritising the right use cases. Most of them are currently internal ones to drive productivity and efficiency or low risk ones to create new revenue lines or improve services for customers.
A long way to go
Lily Haake, Head of Technology and Digital Executive Search at recruitment consultancy Harvey Nash, agrees that AI adoption still has a long way to go. She points to preliminary research for the company’s forthcoming Digital Leadership Report, which is based on a survey of up to 4,000 tech leaders around the world.
These initial figures reveal that only about three out of five organisations are even piloting AI initiatives today. The rest have yet to dabble, although they may consider doing so in future. Moreover, a mere one in five have so far developed any kind of AI policy.
Key barriers to the technology’s adoption include cultural resistance and a lack of access to the right skills (38% respectively). Next in line is being unable to demonstrate a clear business case, followed by not having the right tools and technology in place (31% respectively).
As Haake points out, this situation appears to indicate that “there’s not going to be an AI revolution in the next five years.” The nascent nature of the market is also reflected in the fact that most organizations, unless they are truly serious about the technology, are not yet creating dedicated AI positions internally. They are instead adding AI responsibilities to existing data-related roles. Examples here include:
- AI Architects: In some instances, this is a new role created for an AI specialist and, in others, new responsibilities have been assigned to data architects. Their job is to define and implement AI architecture and both ensure the technology is aligned with business goals and that it integrates with existing systems
- AI Data Engineers: They generally have a data analytics or software engineering background and create and manage the data pipeline, quality and infrastructure
- Prompt Engineers: This in-demand role involves structuring sentences so they can be understood by a generative AI model, enabling it to generate outputs that align with the original intention behind the query. But over time, it is expected to disappear as models become easier to use
- AI Product Managers: While a rarity today, this role bridges the gap between tech and the business, which includes interacting with legal, compliance and diversity, equity and inclusion functions
- AI Business Analysts: Usually business analysts, they have been given additional responsibility for gathering requirements for AI use cases
- AI Ethics Officers: Although this post is still uncommon, many businesses are recognising the importance of addressing ethical issues and so numbers are expected to grow over time. Working closely with AI teams, they are responsible for ensuring models are designed ethically from the outset, for preventing misuse and ensuring the technology adheres to ethical, responsible, sustainability and legal standards.
Another role that is getting a lot of publicity right now though is that of the Chief AI Officer (CAIO). This position is responsible for developing a comprehensive AI strategy that is aligned with the company’s overall business objectives. As such, CIAOs require a mixture of a deep understanding of AI technologies, commercial acumen, strategic vision and good leadership and communication skills.
But again, because so many employers are not yet investing in AI in any serious way, Haake believes there is, in most cases, little need to spend (generally large amounts of) money on this kind of dedicated role:
As AI becomes more pervasive and accessible through the explosion of tools, such as ChatGPT, many organizations feel they require dedicated leadership in this space to effectively harness its potential and align it with their overall business strategy as well as to help them navigate ethical and regulatory concerns. However, this is a very new position and businesses should be cautious of ‘panic hiring’ a role that they may not really need and are not yet ready for. The role is still niche.
Another means of assessing whether a CAIO is required or not is to assess the organization’s data and AI maturity levels. Haake explains:
If data quality is poor, many organizations may first want to get the foundations correct with proper master data management, quality and governance. The readiness for a CAIO depends on the organization’s AI maturity level. Businesses that are at the early stages of AI adoption may not have the necessary infrastructure, data capabilities or AI talent in place to support a dedicated AI leadership role.
Where CAIOs have a valuable role to play
Those employers that could benefit from a CAIO, on the other hand, have usually identified AI as a strategic priority and recognize its potential to transform operations, improve decision-making or boost competitive advantage. They might previously have invested in AI initiatives but not achieved the desired outcomes or operate in a sector that is experiencing rapid AI-generated change or disruption.
Key industries in this bracket include IT, telecoms and ecommerce. Technologically-mature businesses in healthcare, retail, manufacturing and energy may see a need for a CAIO too. As to who the position should report to though, Haake says:
Generally, we would see the CAIO reporting into a Chief Data Officer. The CDO may then report to a Chief Technology Officer or business executive, such as the Chief Financial Officer or Chief Operating Officer. However, if the business is truly digitally-native and perhaps has AI as a major product line, you may see the CAIO on the executive board and/or as a peer to the CTO.
As AI moves increasingly into the mainstream though, other lower-level but new and dedicated roles are expected to appear. These include:
- AI Auditors and Testers: Although it is still very early days for this role, over time it is expected to become as widespread as quality assurance professionals are today. The job will involve establishing whether a system is biased, how to measure its accuracy and also its compliance with internal AI policies and external regulations as they appear
- AI Anti-Bias Specialists: This position is unlikely to emerge until legal action is threatened or scandal hits. But the role would be entail addressing any bias within the algorithm in order to guard against discriminatory outcomes
- AI Co-Pilot: The focus of this role would be training employees on how to use the technology effectively and sharing best practice across the business to ensure better results.
Moreover, other functions beyond IT and digital, ranging from HR to legal, will also likely acquire their own AI specialists to ensure appropriate expertise is embedded where it needs to be.
The second part of this article explores how AI pioneer BT has revamped itself to optimize the effectiveness of its data using AI, the aim being to develop innovative, new commercial offerings as a result.