AI Cybersecurity Pulse

Existing Nigerian statutes can address some deepfake harms but were not designed for synthetic identity manipulation.

• Cybercrimes Act 2015 (amended...

Experts warn AI has already remade cybercrime—making attacks faster, cheaper, and blending fraud with hacking into one operation. One CEO claims this...

An unrevoked master key exposed in May granted attackers owner-level GitHub access in October, allowing them to bypass dual-approval controls, reroute...

AI agents are operating beyond user control across personal and infrastructure domains.

• Banks including JPMorgan and Citi detect third-party agents...

Enterprises face risks from unpredictable AI agents that access data and act without supervision, even without compromise. Microsoft’s expanded...

How effective and independently verifiable are these AI protections against deepfakes?

• OpenAI's election approach forces political image prompts...

The IMF and World Bank’s Bangkok Blueprint reframes cyber-enabled fraud as a systemic policy challenge rather than isolated crimes, demanding...

Healthcare CISOs face budget worries yet cannot ignore $7.4 million breach costs. Summit speakers outlined three practical moves:

• Visibility: Create...

Routine employee interactions with tools like ChatGPT introduce hidden data risks that traditional controls can't fully address.

• Unintentional...

U.S. defense partnerships with Anthropic, Google, OpenAI, and xAI—each worth up to $200 million—reflect a reversal where government now depends on...

Frontier AI developers are urging government regulation amid catastrophic risk warnings, prompting calls for a bipartisan Congressional Commission to...

• Banks struggle to convert individual AI gains into organization-wide efficiency while controlling multi-agent workflows and accountability
  -...

The PHP2Shell emergency fixes for CVE-2026-109618 and CVE-2026-109619 revealed how AI-accelerated discovery narrows the disclosure-to-exploit window...

Responsible AI ethics cover transparency, bias, privacy and accountability, but remain decision frameworks rather than technical mandates.

• Saudi...

AI deepfake impersonation is scaling rapidly, with 54% phishing click-through rates and multimillion-dollar fraud cases hitting organizations of every...

AI systems compile up to 90% of online browsing data to infer personal details like health conditions and income.

This aggregated information lets...

Does your own telemetry deliver earlier, more relevant threat intelligence than paid feeds?

• DNS activity revealed DGA surges in February 2026,...

Two SOC platforms approach trusted autonomy differently, with clear implications for outcomes.

• Splunk ES 8.7 defines trusted autonomy as AI...

Fortinet's new FortiGuard service embeds AI-specific inspection directly into FortiGate firewalls, letting organizations govern AI use through...

• NLP pipeline starts with TF-IDF and n-grams to turn email text into features for Random Forest and XGBoost models
• URL feature extraction adds...