Tighten Contracts and Controls for AI, Identity and Email Security

What happened

SonicWall's 2026 Healthcare Protect Brief shows healthcare remains the most persistently targeted vertical with the smallest year‑over‑year decline in attack volumes. The report flags mass exploitation of remote desktop tools and many unpatchable medical devices that span hospital networks, making this an operational rather than theoretical exposure. Watch whether suppliers managing remote access and device segmentation publish remediation commitments or new managed‑service offerings

Why the category manager should care

Treat healthcare telemetry as an explicit demand driver for hardened access controls and supplier remediation SLAs because attacks are concentrated and persistent

Key facts

• Healthcare recorded the smallest attack decline at just 17% year‑over‑year
• UltraVNC exploitation attempts measured in the millions over five months
• 243 unique IoT attack signatures targeting connected medical devices

What happened

A Five Eyes–aligned advisory frames AI as a force that accelerates existing attack methods rather than inventing wholly new risks. The key operational point is that attackers can automate reconnaissance and lateral movement in hours or days, so identity controls and least‑privilege become primary defenses to slow automated pivoting

Why the category manager should care

Prioritise procurement actions that reduce privilege exposure and formalise supplier obligations on account and access management because automated attacks exploit lax identity controls

Key facts

• Advisory emphasises attackers can operate at 'weeks and in some cases, days' rather than months
• Identity/least‑privilege highlighted as the most scalable defence against AI‑accelerated late

What happened

Proofpoint's analysis of Australian authorised deposit‑taking institutions finds a majority do not use the strict DMARC 'reject' email policy, leaving customer‑facing email vulnerable to spoofing. The study shows a persistent gap between regulator attention on AI-driven threats and basic email controls that materially reduce phishing success

Why the category manager should care

Email‑authentication is a low‑cost, high‑impact control buyers should mandate in onboarding and contract renewals for communication‑dependent vendors

Key facts

• Analysis covered 78 authorised deposit‑taking institutions
• 59% do not use a DMARC reject policy; 41% have adopted reject

What happened

Flux's report finds a large share of organisations are running AI‑generated code in production and many teams have not yet closed oversight gaps. The operational detail to note is that organisations report changes in build/release processes but still struggle to catch security and dependency problems introduced by AI‑assisted code

Why the category manager should care

Treat AI‑assisted code delivery as a deliverable with review and traceability obligations rather than an informal engineering convenience

Key facts

• 44.7% of organisations reported AI‑generated code in production
• Respondents cite security problems and dependency changes as the hardest issues to catch