Business · 19 May 2026
AI in Business
For business leaders and professionals
At askKira, we provide secure, personalised enterprise AI for UK businesses. Our platform prevents data leaks, protects IP and client data, and gives your teams AI-powered tools at a fraction of the cost of other enterprise solutions.
This week's roundup brings you the latest AI developments that matter for UK organisations.
AI Security & Governance
Shadow AI causes one in five corporate data breaches globally
Research shows 97% of companies experiencing an AI breach failed to implement basic access controls. With employees pasting sensitive data into unapproved tools, shadow AI incidents now cost significantly more than standard breaches. Leaders must inventory tools, write simple rules, provide approved alternatives and train staff properly.
Why it matters: UK businesses must move beyond static policies and deploy secure enterprise alternatives like askKira to prevent staff from leaking GDPR-protected data into public models.
Read the source on LinkedIn →Human error drives AI data leaks despite strict corporate policies
Unauthorised AI use is now detected in 73% of organisations but only 28% have technical controls to monitor or block it. Employees routinely bypass lengthy policy documents to meet deadlines by pasting sensitive documents into public chatbots. Security experts argue that enforcement must happen at the browser level rather than relying solely on awareness training.
Why it matters: UK IT directors need to implement endpoint browser controls to stop staff accidentally breaching the Data Protection Act while trying to improve their productivity.
Read the source on LinkedIn →Organisations struggle with AI control rather than technological capability
Many businesses scale AI tools without defining who owns the decisions or what data is trusted. This lack of structure means AI simply adds to manual checks and workarounds rather than streamlining operations. Successful deployments require clear accountability and review processes before any tools reach live workflows.
Why it matters: UK operations directors must establish clear lines of accountability for AI-assisted decisions to ensure compliance with emerging corporate governance standards.
Read the source on LinkedIn →Financial firms fail to embed true accountability in AI governance
Financial institutions frequently confuse having a risk policy with possessing genuine AI governance. Accountability is often assumed rather than assigned, leaving data science and compliance teams pointing fingers at each other. True governance requires weaving risk management into the culture from day one rather than treating it as a final checkpoint.
Why it matters: UK financial services firms must explicitly assign AI risk ownership to meet strict FCA expectations and maintain consumer trust.
Read the source on LinkedIn →AI Regulation
EU AI Act delays high-risk system obligations until late 2027
The European Council and Parliament have agreed to push Annex III high-risk system obligations from August 2026 to December 2027. However, existing rules around prohibited practices and AI literacy remain in force today. Experts warn that businesses should use this 16-month delay to map data flows rather than pausing their compliance efforts.
Why it matters: UK firms operating in Europe must not pause their compliance programmes as UK GDPR and FCA Consumer Duty rules already govern current AI usage.
Read the source on LinkedIn →AI Impact on Work
McKinsey predicts AI will fundamentally reshape human skills by 2030
A new McKinsey report suggests AI could add trillions in economic value by 2030 but will drastically alter how human skills are applied. Workers will shift from executing tasks to orchestrating AI systems, making leadership, communication and empathy more valuable than ever. The focus will move from generating answers to making strategic decisions.
Why it matters: UK employers and educational trusts must urgently adapt their training programmes to focus on strategic thinking and system orchestration rather than rote task execution.
Read the source on LinkedIn →