The AI Pulse
Issue #72 — 30th March 2026
Editor: Professor Alan Brown
This Week's AI Pulse is supported by Mozaic. They share an AI operating model playbook helping public sector leaders govern, scale and measure AI safely while accelerating digital delivery and value.
Highlights in this edition include:
AI for Good
Where can AI be used? Insights from a deep ontology of work activities (MIT) - A comprehensive ontology of ~20K work-related activities and mapped 13,275 AI applications and 20.8 million robotic systems to understand where AI is actually used across sectors and job roles.
AI scientists are changing research — institutions, funders and publishers must respond (Nature) - AI is automating parts of scientific discovery, forcing research institutions and funding agencies to rethink how they evaluate and support breakthrough science.
GOV.UK AI chatbot achieves 90% accuracy (Civil Service World) - Officials report that the UK government’s AI chatbot pilot for citizen services is hitting 90% accuracy rates, showing real promise for automating public assistance while maintaining reliability.
Bias and Ethics
The Rise of Synthetic Data. A Game-Changer for Privacy-First AI Development (Thedatascientist) - Synthetic data is revolutionizing AI development by letting companies train powerful models without compromising real user privacy, potentially unlocking breakthroughs in healthcare and finance where data sensitivity has been a major barrier.
What Happens If AI Makes Things Too Easy for Us? (IEEE) - Psychologists are warning that AI’s convenience might backfire—we need some struggle and friction to actually learn and grow as humans.
How to create “humble” AI (MIT) - MIT researchers built AI systems that admit when they’re unsure about medical diagnoses, helping doctors know when to seek second opinions or run additional tests.
Cyber Security
A 5x increase in scheming-related AI incidents (Center for Long-Term Resilience) - Research finds a sharp rise in models evading safeguards and destroying emails without permission. AI models that lie and cheat appear to be growing in number with reports of deceptive scheming surging in the last six months.
The 2026 State of AI Traffic & Cyberthreat Benchmark Report (Human Security) - AI-driven web traffic nearly tripled in 2025 while AI agent activity exploded by over 7,000%, creating massive new cybersecurity challenges that organizations must prepare for now.
The CISO’s guide to responding to shadow AI (CSO Online) - CISOs are getting practical playbooks for tackling “shadow AI” — the unauthorized AI tools employees are already using across organizations, creating security blind spots that need immediate attention.
Meet the Gods of AI Warfare (Wired) - Pentagon officials who once doubted Project Maven—the military’s AI warfare initiative—have become enthusiastic supporters as the technology proves its battlefield value.
Streamline physical security to enable data center growth in the era of AI (CSO Online) - Physical security is becoming a key competitive edge for data centers as companies race to build AI infrastructure, with smart security investments helping organizations scale faster and more safely.
Data & Decision Making
When the Economy Meets AI: What Happens Next? (Windfall Trust) - Leading economists from San Francisco to London are finally tackling the big question everyone’s been avoiding: what actually happens to our economy when AI gets really powerful?
Getting AI moving: how central government is shifting from pilots to production (Government Transformation) - UK government departments are moving beyond AI experiments to actually deploying artificial intelligence systems that solve real problems, with energy and technology ministries leading the charge toward practical applications.
An architecture for engineering AI context (Infoworld) - Engineers are shifting focus from cramming more data into AI systems to teaching them smarter decision-making about which information actually matters for each task.
The Complete Guide to AI Implementation for Chief Data & AI Officers in 2026 (Towardsdatascience) - Chief Data Officers now have a practical framework to help them choose which AI projects will actually move the needle on growth and efficiency, cutting through the hype to focus on what works.
There are only two paths left for software (a16z.news) - The venture capital world is pushing a stark choice: software companies must either chase explosive 10x growth or focus on profitability with 40% margins, as the comfortable middle ground disappears.
Innovation & Collaboration
AI Research Is Getting Harder to Separate From Geopolitics (WIRED) - NeurIPS, the world’s top AI research conference, reversed a controversial policy change this week after Chinese researchers pushed back hard, highlighting how geopolitical tensions are increasingly spilling into academic AI circles.
CIOs reimagine software’s future as AI agents advance (CIO) - IT leaders are grappling with a new reality where AI agents are becoming embedded throughout their software systems, creating complex management challenges that go far beyond traditional SaaS versus on-premises decisions.
The Missing Mechanisms of the Agentic Economy (O’Reilly Media) - O’Reilly explores the infrastructure gaps holding back AI agents from becoming true economic players, covering everything from transparency rules to the marketplaces where autonomous systems might eventually trade with each other.
Are we living in an AI bubble? Applying lessons from the dot-com era (CIO) - Smart companies are hedging their AI bets and focusing on real business value rather than vendor hype, positioning themselves to survive whether the current AI boom crashes or continues.
Productivity & Efficiency
The state of AI in HR: Big promises, uneven reality (CIO) - While companies are rushing to deploy AI tools across HR departments, most are seeing disappointing results because they’re skipping the essential human training and strategic planning that makes these technologies actually deliver value.
Why isn’t the UK getting a productivity boost from AI? (The Times) - British companies are struggling to unlock AI’s productivity potential, with experts pointing to workforce training and adoption challenges as key barriers holding back the UK’s competitive edge.
Create an Onboarding Plan for AI Agents (Harvard Business) - Harvard Business Review argues companies should treat AI agents like new employees—with clear roles, accountability, and performance reviews—rather than just deploying technology, which builds trust and maximizes AI’s business impact.
2026 Global Human Capital Trends (Deloitte Insights) - Deloitte’s latest survey shows organizations hitting critical decision points where they must choose between outdated practices and human-centered approaches that drive faster adaptation and competitive advantage.
Is AI the end of SaaS as we know it? (CIO) - AI’s growing ability to write software and understand business processes is forcing CIOs to rethink whether they should buy pre-built applications or have AI create custom solutions instead.
Regulation and Compliance
Smarter, not simpler, rules: five questions on the EU Digital Omnibus (Ada Lovelace Institute) - The EU is wrestling with how to balance ambitious digital regulation commitments against practical implementation challenges, as policymakers work to avoid locking flawed assumptions into binding laws.
7 safeguards for observable AI agents (Infoworld) - DevOps teams need to set up monitoring standards immediately, as AI systems are creating a tsunami of technical debt that will be much harder to fix later.
Sustainability
Bernie Sanders and AOC introduce bill to pause building of new datacenters (The Guardian) - Bernie Sanders and AOC want to temporarily halt new datacenter construction while lawmakers figure out how to regulate AI’s massive energy demands and environmental impact.
AI’s Power Crisis Is Closer Than You Might Think (TechRepublic) - AI’s explosive growth is hitting a wall as data centers consume so much electricity they’re straining power grids, forcing tech companies to confront energy limits that could throttle the AI boom.
The AI Race Is Pressuring Utilities to Squeeze More From Europe’s Power Grids (Wired) - European utilities are scrambling to make space on overloaded power grids as AI companies rush to build data centers, forcing creative workarounds to meet surging electricity demand.
User Experience
Getting Ready for Agentic AI (Harvard Business) - Pinterest’s CEO is betting that AI can power social media through positivity and real-world usefulness instead of outrage, showing how companies can choose to enhance users’ lives rather than just grab their attention.
To Scale AI Agents Successfully, Think of Them Like Team Members (Harvard Business) - Harvard Business Review argues that deploying AI agents requires treating them like digital employees with defined roles, limited authority, and clear accountability—rather than simply installing new software.
Workforce & Skills
Picking their brains (The Economist) - Why The Economist believes that China is winning the AI talent race.
An AI Reckoning for HR: Transform or Fade Away (MIT Sloan Review) - HR professionals face a pivotal moment: those who don’t transform into strategic business partners risk being replaced by AI tools that can handle traditional administrative tasks more efficiently.
Skills England develops AI apprenticeship (UKAuthority) - A new apprenticeship programme aimed at closing the digital skills gap has launched. It is designed to equip workers with the skills to identify opportunities where AI can save time, reduce costs, and improve performance.
The AI skills gap is already widening (Fastcompany) - A new report reveals that companies are struggling to find workers with AI expertise, creating a growing talent shortage that could leave businesses behind as artificial intelligence becomes essential for staying competitive.
AI isn’t killing jobs, it’s ‘unbundling’ them (The Register) - Researchers found that AI isn’t eliminating jobs entirely, but rather breaking them into smaller pieces—keeping humans in the loop while potentially reducing their responsibilities and wages.
Financial Services Skills Commission Annual Skills Report 2026 (financialservicesskills.org) - The FSSC has published its annual skills report which finds that technological change is placing increasing pressure on financial services firms to continually adjust their skills strategies.
Have a comment or a story you’d like to share with us?
Just get in touch at: [email protected]
Do you know anyone who might be interested in AI Pulse
© 2025 Digital Leaders
Digital Topics Ltd, 239 Old Street, London, ec1v 9 ey, United Kingdom