Loading story…
Microsoft CEO Satya Nadella warned that enterprises are exposing proprietary knowledge when they use third-party AI systems, arguing that prompts, tool interactions, and corrective feedback can transfer valuable internal expertise to model providers. He described the dynamic as a "reverse information paradox," saying customers may pay twice for AI services—once in fees and again in knowledge that improves a vendor’s models—while urging organizations to set clear trust boundaries, keep control of evaluations and feedback, and avoid dependence on a single provider.
The concern intensified as Google updated Search privacy settings so media from search interactions—including images, files, audio, and video—can be used to develop and improve its AI models unless users opt out. Separate reporting said enterprises are increasingly weighing private or open models against closed AI services because of fears that sensitive business data could be retained, learned from, or used to strengthen vendors that may later compete with them, even as major providers maintain that some API traffic is excluded from training or covered by zero-retention options.
Track how attackers are adapting to this technology.
2 events from the most recent confirmed update back to the earliest known activity.
Microsoft CEO Satya Nadella said enterprises risk giving AI providers proprietary knowledge through prompts, tool usage, and corrective feedback, describing it as a 'reverse information paradox.' He called for stronger protections, trust boundaries, and retaining control over evaluations and feedback.
Google updated its Search privacy settings so that media saved from interactions with Google Search services, including images, files, audio, and video, may be used to develop and improve its AI models and related services. The change reportedly auto-enrolled users unless they manually opted out through privacy settings.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
3 references tracked. Mallory keeps watching after this page renders.
thenewstack.io
Open sourceitpro.com
Open sourcezdnet.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Meta suspended its internal Model Capability Initiative after an internal error reportedly made sensitive employee monitoring data broadly accessible inside the company, intensifying criticism of a program that collected workplace activity from tools including Gmail, Google Chat, Meta AI, screenshots, and usage patterns. Meta said it found no evidence of malicious activity or an external breach, but the incident renewed concerns about consent, unencrypted storage, and how far companies are going to feed AI systems with employee data. Similar privacy concerns surfaced in the Netherlands, where a watchdog said the AIVD and MIVD intelligence services mishandled bulk datasets containing personal information through weak controls, excessive retention, and inadequate oversight, raising alarms that AI-assisted analysis could be applied to unlawfully sourced or improperly retained citizen data. At the same time, new reporting and vendor changes underscored how quickly enterprise AI is expanding the attack surface and governance burden. Orca Security said **81.2%** of organizations running AI packages had at least one known vulnerability and that **99.9%** of fixable AI vulnerability alerts remained unpatched, with risks spanning agent frameworks, vector databases, exposed API keys, excessive permissions, weak authentication, and internet-facing services such as Amazon SageMaker, Azure OpenAI, and Google Vertex AI. Microsoft is adding an in-meeting Teams control to disable Copilot, Facilitator, and recap features during sensitive discussions, while broader industry debate has focused on employees oversharing data with AI assistants, the need to govern agent identities and registries, and warnings that AI prompts, memory, and orchestration layers can leak or concentrate proprietary organizational knowledge if enterprises do not retain tight control over access, telemetry, and data handling.
Jul 16, 2026
A Lawfare analysis identifies **replacement through knowledge acquisition (RKA)** as an enterprise risk in which frontier AI providers derive valuable knowledge from customer prompts, uploaded files, outputs, metadata, and usage patterns, then use it to develop products that compete with those customers. The concern spans software, legal services, financial services, and potentially manufacturing, particularly as AI platforms expand across large business customer bases. The analysis warns that data acquisition could occur openly through permissive contracts or more subtly through retained enterprise data, continual learning from transient interactions, and commercial pressure to accept broad sharing terms. It argues that existing contracts, trade-secret protections, and U.S. antitrust law may not adequately prevent such misuse, and recommends stronger transparency and audit requirements, safeguards for contractual information sharing, duties restricting customer-data misuse, and statutory damages.
Aug 26, 2026
The rapid adoption of generative AI chatbots, such as ChatGPT, is transforming both consumer and enterprise environments, with significant growth in usage and market value. These chatbots are being used for a wide range of applications, from customer service to code generation and mental health support. However, their increasing prevalence raises concerns about risks such as hallucinations, dangerous suggestions, and the need for robust guardrails to ensure safe deployment and use. Simultaneously, privacy concerns have emerged regarding how major technology companies, like Google, may use personal data to train AI models. Google recently denied allegations that it analyzes private Gmail content to train its Gemini AI model, following a class action lawsuit and public confusion over changes in Gmail's smart features settings. The company clarified that while smart features have existed for years, Gmail content is not used for AI model training, and any changes to terms or policies would be communicated transparently. These developments highlight the ongoing tension between AI innovation, user privacy, and the need for clear communication about data usage.
Jun 29, 2026