Insights

20/8/2026

The NCSC's Q2 2026 report shows overall incident reports held steady at 1,129, but those needing specialist technical support for potential national significance rose from 77 to 92. Financial losses fell 52 percent to $2.7 million, with scams and fraud most common, and the report flags rising malware and QR code phishing scams, including recent incidents involving Christchurch parking meters.

4/8/2026

MYOB's 2026 Business Monitor found 36 percent of New Zealand SMEs now use AI, up from 32 percent last year, with 23 percent using it for HR and recruitment support. Adoption is uneven though, ranging from 30 percent for businesses with one to five staff up to 64 percent for those with 20 or more, and medium-sized firms are the most reliant, with 92 percent saying they'd feel the impact immediately if they stopped.

4/8/2026

Microsoft is rolling out a "Report a concern" feature in Teams that lets participants flag suspicious activity, such as phishing, impersonation or AI deepfakes, directly from a meeting. Reports go to the Teams admin center and Microsoft Defender portal for IT teams to investigate, part of a wider response to rising AI-enabled meeting fraud, and it's expected to be fully available by the end of August 2026.

31/7/2026

Microsoft's July 2026 Copilot update adds agents in Copilot Chat, new features across Word, Excel, PowerPoint and Teams, an access heatmap and AI summaries for SharePoint and OneDrive, more Copilot connectors, and new admin controls and reports.

28/7/2026

Health New Zealand has launched a new national framework, the AI Pre-Implementation Evaluation Framework, to standardise how AI tools are assessed before they're used in the health system. Developed over 18 months by Dr Rosie Dobson and endorsed by the Health NZ Board, it ensures AI tools are only adopted where they're safe, ethical and genuinely benefit patients and staff.

1/7/2026

A cyber security expert is warning New Zealand isn't ready for a national cyber emergency, after three C2 level incidents (the second-highest severity) hit organisations of national significance this year, including breaches at Manage My Health, MediMap and IntraCare. Aura Information Security's Patrick Sharp says many businesses still lack basics like multi-factor authentication, and only half of boards surveyed have even discussed cyber security or tested their response plans.

23/6/2026

The Five Eyes intelligence alliance has warned that advanced AI models are improving fast enough to outpace current cybersecurity defences within months rather than years, urging organisations to update old systems, integrate AI into their own defences, and tighten access to critical systems. The warning follows Anthropic's disclosure that its frontier models can find software vulnerabilities with unprecedented skill, with a former US cybersecurity chief calling the alert genuinely alarming.

3/6/2026

GitHub has moved Copilot from flat-rate requests to usage-based pricing tied to tokens, and many developers are only now realising how much their everyday habits cost. Some are reporting a handful of prompts burning through hundreds or even thousands of credits, prompting a rethink of how often they lean on AI and, for some, a look at cheaper alternatives.

1/6/2026

New WEF-linked research finds a widening gap between executive enthusiasm for AI and worker unease, identifying five distinct employee postures towards adoption, from enthusiasts and the curious through to cautious, sceptical, and openly opposed staff. Even employees who publicly comply with AI mandates often quietly resist the tools behind the scenes, and the researchers argue successful adoption depends more on trust and open communication than on the technology itself.