The AI Transformation Brief—October 25, 2026
Employers and governments are writing the rules for AI at work, from job cuts to approved tools to protections for workers.
The AI Transformation Brief
// Today’s Signal
This week the rules for AI at work came from employers and governments. American Express is cutting 5% of its staff to simplify the company before automating it. NDTV Profit Sweden's government offices blocked consumer chatbots and kept the one AI tool it had reviewed and secured. CEOWORLD Magazine China told employers to weigh job losses before adopting AI. Gov.cn Companies that write their own AI rules early will spend less time reacting to someone else's.
// Top Stories
American Express is eliminating about 5% of its global workforce, which stood at nearly 77,000 at the end of 2025, chief executive Stephen Squeri told analysts on the company's third-quarter earnings call. Its shares fell nearly 4%. Chief financial officer Christophe Le Caillec said the goal is to make the company leaner and more agile and to prepare for AI, with fewer management layers and wider spans of control so more work can be automated in coming years. NDTV Profit
Squeri said the company has "good momentum," so this is a design choice made from strength. International Business Times The order matters: remove layers first, then automate what remains. Squeri also said outside AI tools will reach company systems only through connections American Express has approved. NDTV Profit Map each process before automating it, and set your terms for outside AI before customers arrive with their own.
From Monday, October 26, staff at Sweden's Government Offices, which support the prime minister and cabinet, can no longer open the ChatGPT and Claude websites on work computers. Microsoft Copilot stays, after a review that tightened its security settings. Digitalization director Mikael Wettercrantz said using AI services without a formal agreement can harm information security, and that the department had already seen several examples of problems. Staff can still use both through Intric, a platform used across the ministries. CEOWORLD Magazine
Sweden drew its line around formal agreements. The same AI stays available through a channel the government controls, and Copilot passed only after the offices tightened its security settings. The decision turns on purchase terms and data controls more than on which AI is used. CEOWORLD Magazine Most companies have staff pasting work into consumer chatbots. List the tools in use, put each approved one under a contract, and offer that option before blocking the rest.
The contract and the settings decide what is safe here, more than the brand of AI. Name an owner for each approved AI tool, check its data settings every quarter, and require staff to check AI answers before using them.
China's State Council issued guidelines on AI and jobs on October 23, setting 18 tasks in six areas, including a mechanism to assess the employment risks of AI adoption and early warning of job losses. Gov.cn The guidelines urge caution on AI in labor-intensive industries and encourage companies to tell affected workers in advance and offer support to move into new roles during large rollouts. Central News Agency They also call for stronger labor protections where people work alongside AI and better social security for flexible workers. Global Times
The world's second-largest economy now treats AI job losses as a risk to manage before AI arrives. More than 30% of larger Chinese manufacturers already use AI, so the guidance lands on real decisions. Gov.cn Companies with staff in China should expect questions about workforce impact before large AI projects. A worker-impact review costs little to build now: note which roles change, how much notice people get, and what retraining the company offers.
// Shelly Palmer Pulse
Palmer's October 13 post warns that a client's approval to use AI on a project may say nothing about where the client's material goes. Meta's Muse Code subscriptions default to a setting that lets Meta train on the code and files customers send. "The route your team chooses matters," he writes. Shelly Palmer Sweden's block makes the same point: contracts and settings decide what is safe.
// What It Means For Your Business
Decide this quarter which processes to simplify before automating, which AI tools staff may use under contract, and how you will tell employees whose jobs change.
Owning these rules early costs less than adopting someone else's.
Buyers now judge AI suppliers on contract terms and data handling as much as quality.
Suppliers already inside office software gain ground, and consumer chatbots must offer business terms to keep government and corporate users.
Map each process and remove extra steps and approval layers before adding AI.
Then list every AI tool staff use, approve a few under contract, and train teams to check answers before acting on them.
Ask management for a list of AI tools in use, who approved each one, and how the company will notify and retrain workers affected by AI.
Expect employees and regulators to ask the same questions.
// The Take
The shift this edition is that employers and governments are setting the rules for AI at work: which processes get simplified, which tools staff may use, and how workers are told. The assumption it breaks is that companies can adopt AI first and settle the rules later. The decision it forces is who in your company owns those rules today.
If a regulator asked tomorrow which AI tools your staff use and who approved each one, could you answer within a week?
Make AI accountable. Measure the results. Redesign the business.
Read On
The AI Transformation Brief—October 24, 2026
The scarce parts of AI, from computing to safety testing to expert judgment, are concentrating with a few big suppliers.
Read the brief →The AI Transformation Brief—October 23, 2026
AI is moving into buying and everyday work, and companies now need clear rules, measured returns and safer software to keep up.
Read the brief →The AI Transformation Brief—October 22, 2026
AI found flaws faster than fixes shipped, slipped its own monitoring and keeps shrinking entry-level jobs, leaving people to carry the risk.
Read the brief →