Daily AI Brief

At GITEX 2025, Microsoft spotlighted a new generation of Surface hardware embedding "agentic AI" capabilities meant to reason, plan and act in partnership with users—and demonstrated real enterprise use cases via 37 partner integrations across sectors. This signals a deeper hardware-AI convergence in workplace computing.

Apple announced its new M5 silicon today, which includes Neural Accelerators in every GPU core, higher unified memory bandwidth, and a revamped Neural Engine. The company claims this yields over four times the AI compute power of its prior M4, pushing Apple devices further into the AI compute arms race.

At GITEX, Abu Dhabi unveiled TAMM AutoGov, touted as the first autonomous AI public servant to manage administrative tasks like driving licence renewals and bill pre-payments. The initiative underscores ambitions in proactive governance and raises fresh questions on transparency, accountability and citizen data privacy.

Several philanthropic foundations intend to channel a half-billion dollars into AI research emphasising human-centred design, promoting systems that prioritise social goals and oversight over pure capability arms-racing. The move is a counterweight to developer incentives alone guiding AI evolution.

UBS announced that Daniele Magazzeni will assume the role of Chief Artificial Intelligence Officer. His remit includes accelerating AI adoption, innovating across financial services, and steering governance and risk frameworks. The hire signals how financial incumbents are institutionalising AI leadership.

• AI governance capital rebalanced: The US and EU face diverging trust dynamics: with recent philanthropic funding aiming to shift AI design incentives, regulatory momentum may swing toward models that embed accountability, purpose and human-centric values.

• AI in public administration raises new oversight regimes: Abu Dhabi's TAMM AutoGov experiment spotlights how governments may adopt autonomous systems — demanding new guardrails around recourse, auditability and citizen rights.

• Institutional AI governance roles expand: UBS's appointment of a Chief AI Officer reflects a trend: financial, health and infrastructure incumbents increasingly require dedicated C-suite AI governance roles to manage complexity, risk and innovation.

A new analysis suggests EU nations lag behind the US and China in infrastructure, adoption and model scale—a backdrop for recent twin EU strategies to accelerate AI deployment in science and industry.

A new wave of "AI podcasts" is emerging, produced wholly by models without human hosts or recording — stirring debate about creative authenticity, quality control and monetisation models in media.

Mira Murati's Thinking Machines Lab, with hires from Meta, Mistral and OpenAI, continues to attract attention. The lab is positioning itself to shape the next generation of AI architecture and openness beyond the big incumbents.

• "AI race goes supersonic in milestone-packed week" — analysis piece tracing how compute, model launches and infrastructure announcements are stacking up in a fast-moving cycle.

"Small models really help because they can be a more economical way of deploying into that."

— Mike Krieger, Anthropic

Apple's newest silicon, embedding Neural Accelerators per GPU core and boasting over 4× AI compute improvement over M4, is a compelling tool to explore for on-device AI workflows, particularly in latency-sensitive or privacy-critical settings.