The scale of AI adoption makes AI governance increasingly relevant, if not a strategic necessity. McKinsey’s report The State of AI in 2026: On the Road to ROI reports that nearly nine in ten survey respondents say their organizations regularly use AI in at least one business function, while 56% report use in three or more. Organizations are also moving from the experimentation phase towards scaling, and organizations are adopting AI agents at a growing pace.
Yet adoption figures tell only part of the story. The AI transformation is happening and it has huge impacts. However, the impacts are spread out unevenly across different organizations. To understand those differences, we need to look at how AI relates to an organization’s core business. The significance of AI depends on its relationship to the organization’s core work, assets, and sources of revenue.
Consider a local plumbing business. It might use generative AI to draft customer emails, prepare marketing material, or help organize information for quotations. These applications can improve its operations, but its main service still involves diagnosing faults and carrying out physical repairs. The core elements of its business remain unchallenged.
Is AI use supporting the business or changing its products and business models entirely?
For a translation agency, the relationship is different. AI can produce the very output the customer is purchasing. This can shift demand from translating an entire document towards reviewing, correcting, and taking responsibility for a machine-generated translation. It can also change delivery times, pricing, staffing, and the skills clients value.
A publisher may use the same tools as the local plumber or hairdresser for similar administrative purposes. However, AI can also generate text that competes with original works, summarize publications, and distribute answers drawn from published material. It therefore affects the publisher’s core product and the ways readers discover and consume it. An AI-generated answer could help someone find a publication or reduce their need to visit the publisher’s website. It is likely to be the latter, unfortunately.
Compare a gardening company with an illustration agency. Both can use AI to write proposals, manage enquiries, and promote their services. But generating an image of a garden does not maintain the customer’s garden. Generating an illustration can, in some circumstances, provide an alternative to commissioning one.
Physical services can also change through robotics, customer expectations, and competition. But their core tasks are generally less directly exposed to generative AI than work involving digital content.
For the illustration agency, AI therefore affects more than efficiency. It can change which assignments clients commission, how much they are willing to pay, and what they expect an illustrator to deliver. Some may prioritize speed and low cost; others may place greater value on artistic judgment, a distinctive style, or a close working relationship.
These are illustrative comparisons, rather than claims that particular businesses are unaffected. Physical services can also change through robotics, customer expectations, and competition. But their core tasks are generally less directly exposed to generative AI than work involving digital content. Occupational exposure research provides a useful basis for this distinction.
For rightsholders of creative works the impact of AI is a key topic currently. Their works are used to train and develop AI models, often without explicit approval and without any compensation. AI is then used to create content that competes with the original works.
Publishers face an several layers of changes and challenges. Their content is a product for readers and can also be an input into AI systems. A book publisher may use AI to support editing, translation, or marketing. At the same time, books may be sought for AI training or other uses. This creates licensing opportunities alongside questions about permission, compensation, and the publisher’s authority to license particular rights.
A news publisher faces changes in distribution as well. AI-generated summaries and answers may help people discover its reporting, but they may also satisfy a reader’s immediate information need without a visit to its website or purchase of a tangible copy of a work. That possibility matters to a business whose revenue depends on subscriptions, advertising, or direct relationships with readers.
Unlike the plumbing business, the publisher must consider AI’s effects on production, the reuse of its assets, and access to its customers.
A photographer’s image, for example, may be used to train a system that enables customers to generate images instead of licensing photographs. An author’s books may contribute to an AI model that produces text. Outputs from that AI model may reproduce the works that were used to train the model or they may compete with original works on the market.
For rightsholders – whether individuals or organizations – there are several interests at stake. Are they able to retain control over how their works are being used and by whom? This may involve e.g. prohibiting the use of their works or permitting it through licensing. Can they establish which works have been used, by whom, and for what purposes? Transparency is of key importance. Does the use of their works for AI training and development create a source of income or are their works used withour any remuneration to the authors? And finally, whether the AI-generated content saturates the market so that it negatively affects the sales of original works.
AI governance has particular weight in the creative sectors.
AI governance has particular weight in the creative sectors. Organizations need clear policies for their own use of AI, but they also need a position on how their works are used by others. Internal decisions about tools, quality, and accountability connect with external decisions about licensing, rights protection, and industry standards. The two spheres are interlinked, and the elements of good governance, i.e. proper governance structures and decision-making processes, transparency and accountability are vital for both.