1. Citi Acquires AI-Powered Rewards Fintech Kard to Personalise Card Offers
FinTech Futures · Finance & Banking
Citi’s US Consumer Cards unit has agreed to acquire Kard Financial, a rewards and commerce media platform that uses machine learning to connect banks and merchants through personalised, transaction-data-driven offers. The deal will give Citi’s 70 million cardholders more tailored credit card rewards by combining the bank’s payments scale with Kard’s predictive AI and merchant relationships. Terms were not disclosed, and both companies will continue to operate independently until the acquisition closes.
https://www.fintechfutures.com/m-a/citi-to-acquire-kard-financial
2. Pro Se Litigant Hides Secret AI Instructions in Court Filing
Above the Law · Law & Legal
A court employee in Connecticut discovered hidden white-on-white text embedded in a court filing by a pro se plaintiff — invisible to human readers but legible to AI software. The litigant had smuggled covert AI prompts into the document in an apparent attempt to influence any AI tools used to review it. The incident highlights a novel and alarming frontier in AI misuse within the legal system.
3. MHRA Narrows AI Ambient Voice Tech Rules, NHS Adopts Guidance Immediately
8foldgovernance.com / MHRA · Healthcare & Pharmaceutical
The UK’s Medicines and Healthcare products Regulatory Agency (MHRA) published new guidance on 29 July 2026 clarifying that AI tools which only transcribe or summarise clinical conversations — where a clinician reviews the output — are not regulated as medical devices. NHS England swiftly adopted the guidance, reversing its much broader stance from April 2025. The move brings much-needed regulatory clarity for AI ‘ambient scribe’ products widely used across the NHS.
https://8foldgovernance.com/mhra-ambient-voice-technology-guidance-ai-scribes-nhs/
4. Meta Expands AI Ad MCP Server Letting Any Agent Manage Campaigns
AdMake AI Blog · Media & Marketing
Meta has expanded its first-party MCP (Model Context Protocol) server at mcp.facebook.com/ads, enabling any AI agent to read reporting, create campaigns, and manage product catalogs through a single sanctioned endpoint. The update also introduced MCP server governance rules, letting business portfolio owners control exactly what AI agents can do — from budget changes to catalog updates. Agencies are advised to begin the App Review process now before client demand surges in Q4.
5. Cyabra Wins AI Contract to Help Asia-Pacific Intelligence Agency Counter Disinformation
Military Embedded Systems · Defense & Security
Cybersecurity firm Cyabra has secured a multiyear, six-figure contract to provide an unnamed Asia-Pacific intelligence agency with an AI-powered platform designed to monitor online information and detect coordinated influence operations. The deal expands Cyabra’s growing footprint with government and defence customers across the region. The move reflects surging demand among national security agencies for AI-driven tools to identify and disrupt foreign disinformation campaigns in real time.
6. Microsoft Launches MAI-Cyber-1-Flash AI Model to Automate Vulnerability Remediation
ExploreSec AI Cybersecurity Newsletter · AI Cybersecurity
Microsoft has announced MAI-Cyber-1-Flash, a cybersecurity-focused AI model built to identify and remediate software vulnerabilities within its MDASH security platform, claiming it can handle up to 90% of vulnerability analysis tasks while cutting operational costs by 50%. Alongside it, Microsoft introduced Project Perception, an agent-based security platform using specialised red, blue, and green team AI agents to continuously detect, investigate, and remediate threats while keeping humans in control. Microsoft and NVIDIA have also co-founded the Open Secure AI Alliance with dozens of other organisations to develop open-source AI security standards and tools.
https://www.exploresec.com/blog/2026/8/8/august-2026-exploresec-ai-cybersecurity-newsletter