Field report NE-62 / K7 · sheet /boards

Boards & GovernanceFiled from the territory

IBM warns British firms cannot prove their AI is under control

Only 27% of executives can prove to auditors that AI is under control, IBM says, despite 87% claiming confidence as British firms face sovereignty and compliance risks.

Grid
NE-62 / K7
Surveyed
November 17, 2026
Reading
3 min read
Correspondent
Amara Osei
IBM sounds alarm over AI governance at British firms
Fig. NE-62 / K7IBM sounds alarm over AI governance at British firms — MagicDesk / pixabay

Field notes

  1. 87% of executives claim their AI is under control, but only 27% can prove it to an auditor, according to IBM’s Digital Sovereignty Index.
  2. Most British businesses have faced at least one sovereignty challenge during the previous three years, IBM said.
  3. IBM UK and Ireland chief executive Leon Butler says governance must be built into business processes from the start.
  4. London Stock Exchange Group executive Debbie Lawrence says evidence must be assembled proactively and remain available when an incident occurs.
  5. Certain data sets are worth more than five years ago because companies can now identify historical trends and patterns more easily and quickly.

Only 27% of executives can prove to an auditor that their artificial intelligence systems are under control, even though 87% say they are, according to IBM’s Digital Sovereignty Index. IBM UK and Ireland chief executive Leon Butler warned that British companies risk future compliance issues as governance lags behind the rollout of advanced technologies.

IBM’s recommendation is to build governance into core business processes from the start, rather than add controls after deployment. Butler described the approach through an AI operating model (a framework linking process outcomes, redesign, data and governance), stating:

“We have something called an AI operating model, which is looking at the outcome, the process, and redefining that process, and then making sure that the right data and governance follow it…[if] you build it up, that governance will follow at an overall level.”

The model starts with an intended business outcome, then examines and redefines the process around it. Data and governance follow that redesign, creating a sequence in which control measures accompany operational change rather than arrive later as a compliance exercise.

What must British firms prove?

The 60-percentage-point gap between executive confidence and auditability defines the governance deficit. IBM’s index records that 87% of executives claimed AI was under control, while only 27% could prove that position to an auditor.

Digital sovereignty concerns control over data and technology. IBM, the New York-listed technology group, said most British businesses had faced at least one sovereignty challenge during the previous three years.

Butler placed stronger evidence at the centre of digital sovereignty strategy. His prescription links tighter evidence to greater strategic control, rather than treating governance as a separate compliance function.

At London Stock Exchange Group, Debbie Lawrence, group head of data strategy and management, said evidence must exist before an incident, not after it. Having systems and tools is insufficient unless leaders identify the required proof proactively; when an event occurs, there is no time to scramble.

Her standard was operational: the necessary evidence must be immediately available when an issue arises. That requirement moves compliance work upstream, from retrospective reconstruction to advance preparation, and makes governance evidence part of operating readiness rather than an administrative exercise.

Three priorities follow directly from Butler and Lawrence’s comments:

  • Start with the intended outcome of each business process.
  • Attach the appropriate data and governance to the redesigned process.
  • Assemble audit evidence proactively so that it remains available during an incident.

Why is data becoming more valuable?

Lawrence said more companies now recognise data as an asset and increasingly focus on authenticated data (data whose origin and integrity can be verified). She argued that treating data as a corporate asset will become an increasingly consequential strategic discussion.

“If I look now across the industry, more and more companies are recognising that their data is an asset,” Lawrence said. “The concept of data as an asset is going to become a really interesting discussion because I think people are going to recognise that data is valuable [and the] crown jewels [of a business].”

Lawrence also believes certain data sets are worth more today than they were five years ago. Improved data mining (the extraction of patterns from information) lets companies examine history more easily and faster than they could five years ago. Analysts can identify trends and patterns quicker, increasing the value of records that previously offered less accessible insight.

The strategic decision is clear: British firms must make governance evidence part of AI deployment, not an afterthought assembled after an incident. IBM’s 60-percentage-point gap between confidence and proof and the three-year record of sovereignty challenges leave little room for retrospective compliance.

via cityam.com (Source)

Share this article:

More from Amara Osei

Amara Osei

Show full bio

Market editor covering people, launches and day-to-day practice in corporate strategy and chief-executive leadership at CEO Frontier.

449 articles

Filed under

Adjacent sheets

« Previous articleNext article »

End of sheetNE-62 / K7 · November 17, 2026