This time around, we cover the importance of people above tools, some developments in AI, and the world’s first underwater data center. Reminding us, if nothing else, that IT’s frontier is best placed in capable, human hands.
People-first transformation drives tech success
A new McKinsey report finds that people-first operating models are often the key to successful tech transformations. With around two-thirds of leaders saying their organizations are overly complex and inefficient, leaders who prioritize culture, upskilling, and employee experience get more value from AI and digital tools than those chasing new technologies purely for their own sake.
Unapproved AI tools flood the workplace
Omnissa’s State of Digital Workspace 2026 shows just how quickly AI tools are spreading across enterprise environments, with usage of AI assistants growing by nearly 1,000% across managed devices. Usage has surged across these environments, often faster than policy, governance, or IT oversight can keep up.
As well as a compliance headache, shadow AI is a support, security, and visibility issue across the digital workplace. Because if teams are using tools IT cannot see, let alone manage, the risk is very real.
AI models shelved due to export ban
Not long ago, Anthropic told us that their latest model, Mythos, was too powerful for certain tasks, only to release it to the public, under extremely strict guardrails, as Fable 5. Three days later on June 9, the American government seemed to concur with Anthropic’s view by imposing an export control ban on Fable 5 and its unrestricted version, Mythos 5, citing security concerns. This has caused Anthropic to cut off access to both models worldwide.
The problem, according to security expert Katie Moussouris, is the jailbreak cited is a capability shown by numerous other models and used every day by cybersecurity defenders. Removing it means removing a good portion of its usefulness for cybersecurity defense.
Read more: The Fable 5 Export Controls Harm US Cyber Defense↗
LLMs mirror human trust behavior, study finds
A recent Royal Society paper explored how large language models “trust” humans across five different decision scenarios. The study, which analyzed 43,200 simulations across five LLMs, found patterns that broadly resemble human trust behavior, but also more rigid responses and stronger systematic bias in some cases.
That matters as AI moves into triage, support, and recommendation workflows. Because a calm tone and polished output can give systems an air of objectivity they have not earned. Human oversight still matters, especially where decisions impact access, priority, or user experience.
Read more: A closer look at how large language models ‘trust’ humans: patterns and biases↗
Underwater data center goes live, runs on wind power
China just switched on the world’s first underwater data center, powered mainly by offshore wind turbines. Using seawater for cooling and saving over 90% of land use, the facility offers a really cool glimpse of how future digital infrastructure might become more sustainable.
Read more: World’s first underwater data center is now online↗
Conclusion
Across every headline, one thing sticks out: people, trust, and oversight matter more than ever. So, keep the human element front and center as IT evolves—and you’ll be better prepared for what’s next. Until next time!
Explore more insights
-
Why unifying your IT platform is the most practical way to reduce risk, prevent audit failures, and restore confidence
-
Discover 5 essential features to overcome remote IT security challenges. Learn how your IT security team can work remotely while keeping your business protected.
-
Our Chief Commercial Officer, Peter Turner, examines how remote technology is bridging generational gaps in the modern workforce.
