Tags: data security

Meta's Superintelligence Labs rolled out Muse Spark, a new generative AI model that can analyze users' personal health information, through the Meta AI app. The company says the tool was trained with input from more than 1,000 physicians and will soon appear on Facebook, Instagram and WhatsApp. Health experts warn that the service is not HIPAA‑compliant, may retain data for future training and could expose sensitive information, raising serious privacy and safety questions. Ler mais

A recent Malwarebytes survey reveals that a large majority of respondents are uneasy about artificial‑intelligence tools using their data without consent. Nearly nine out of ten worry about AI privacy, and a similar share avoid sharing personal information with ChatGPT or Gemini. As a result, over forty percent have stopped using each chatbot. The same respondents are also pulling back from social platforms like Instagram and Facebook, while adopting privacy measures such as ad blockers, VPNs, and opting out of data collection. Ler mais

Meta's Ray‑Ban smart glasses, praised for their camera and audio capabilities, are drawing criticism for their privacy practices. When users invoke AI features, the company may send captured media to the cloud, where third‑party contractors could review it to improve services. Meta asserts that non‑AI photos and videos remain on the device unless users opt into cloud storage, but the definition of that storage and the safeguards around it remain vague. The lack of clear encryption and detailed guardrails has left users uneasy about the potential exposure of sensitive personal information. Ler mais

A new Nexos.ai study finds that while 70% of legal workers are already using general‑purpose AI tools, 43% of organizations have no formal AI policies or plans to create them. The biggest risk for small‑ and medium‑sized businesses (SMBs) is not reckless AI use but invisible workflow changes that leave sensitive data exposed. Researchers urge SMBs to adopt simple, clear policies that define approved tools, restrict sensitive data, and require human oversight before AI‑generated content is used in legal work. Ler mais

Researchers demonstrated that large language model (LLM) agents can extract identity clues from free‑text data, search the web autonomously, and match those clues to real‑world individuals. In experiments using interview transcripts, Reddit comments, and a large pool of Reddit users, the AI was able to correctly re‑identify a measurable share of participants while maintaining high precision. The findings highlight a growing capability of AI to breach pseudonymity, raising concerns about privacy in online platforms. Ler mais

Organizations adopting AI face a paradox: powerful AI models need massive compute, yet the hyperscalers that provide it often cannot guarantee that sensitive data remains protected or compliant. To resolve this tension, many are turning to sovereign‑first cloud architectures, hybrid and multi‑cloud deployments, and zero‑copy designs that keep data within national borders and under strict control. These approaches aim to balance the need for scalable AI compute with evolving privacy regulations and the emerging security challenges of AI agents. Ler mais

Sen. Elizabeth Warren (D-MA) has written to Google CEO Sundar Pichai asking for details on the new checkout feature in the Gemini AI chatbot. She warns that the integration could let Google and retailers exploit sensitive user data or push consumers toward higher‑priced items. Warren seeks clarification on what data will be shared with retailers, how pricing might be affected, and whether users will be told when product suggestions are driven by upselling or advertising motives. Google has until mid‑February to respond. Ler mais

The acting head of the Cybersecurity and Infrastructure Security Agency (CISA) uploaded internal government documents marked “for official use only” to the public ChatGPT platform, triggering automated security warnings. The director, Madhu Gottumukkala, had previously received an exception to use the tool, despite a department-wide ban. Homeland Security officials are assessing potential security impacts, while a CISA spokesperson described the usage as short‑term and limited. The incident raises concerns about the handling of unclassified but sensitive data on public AI services. Ler mais

AI‑first browsers embed generative tools such as summarization, rewriting and real‑time suggestions directly into the web‑page experience. While they boost productivity, they also blur the line between approved enterprise software and shadow AI, making it harder for organizations to see when employees invoke AI and what data is processed. This hidden usage creates version drift, skips formal review steps, and shifts interpretation away from source documents, leading to gaps in audit trails, retention, compliance and operational consistency. Experts recommend new controls to keep AI‑generated content traceable and governed within existing workflows. Ler mais

Anthropic introduced Cowork, a new capability for its Claude AI that lets subscribers grant the chatbot access to a MacOS folder. Users can chat with Claude to organize files, rename items, and generate spreadsheets or documents from the folder's contents. The feature, currently limited to Claude Max subscribers at $100 per month, also links to connectors for app integration and works with the Claude Chrome extension. Anthropic cautions that Cowork is in a research preview, recommending use only on non‑sensitive data and noting defenses against prompt‑injection attacks. Ler mais

OpenAI introduced ChatGPT Health, an AI‑driven virtual clinic that can read electronic medical records and fitness data to offer personalized health advice. While the service promises clearer explanations of medical jargon and quicker insight into test results, experts and users voice strong concerns about data privacy, the lack of HIPAA coverage, and the risk of AI hallucinations. Trust, transparency, and regulatory safeguards are cited as essential before widespread adoption can be considered safe. Ler mais

OpenAI announced a new ChatGPT Health tab designed to handle medical questions in a dedicated, private space. The feature separates health chat history, offers encryption, multifactor authentication, and promises that health conversations will not be used to train the model. Users can link wellness apps such as Apple Health and MyFitnessPal. While the service is not intended for diagnosis or treatment, experts warn that the lack of HIPAA coverage could expose health data to inadequate protections. OpenAI says the tab is currently in beta and invites users to join a waitlist. Ler mais

OpenAI has launched ChatGPT Health, a new section within the ChatGPT app designed specifically for health‑related queries. Users can securely link medical records and wellness apps such as Apple Health, MyFitnessPal, and Peloton, allowing the AI to tailor responses to personal data. The tool is positioned as a support system rather than a diagnostic service, emphasizing that it should not replace professional medical care. OpenAI highlights extensive physician collaboration, layered security measures, and the ability for users to control data access and deletion. Access is currently limited to a waitlist, with broader rollout planned in the coming weeks. Ler mais

OpenAI has introduced ChatGPT Health, a separate space within its AI chatbot that lets users link medical records and wellness apps for more personalized health‑related answers. The company says the feature includes extra privacy safeguards and that conversations in this area will not be used to train its foundational models. Still in testing, the service has regional limits on which health apps can connect. OpenAI stresses that ChatGPT Health is not meant for diagnosis or treatment and warns that AI chatbots are not qualified to give medical advice, citing risks of inaccurate information and privacy concerns. Ler mais

Google is unveiling a new cloud‑based platform called Private AI Compute that lets users access more advanced artificial‑intelligence features while keeping their data private. The service mirrors Apple’s Private Cloud Compute by keeping sensitive information visible only to the user, even from Google, and by moving heavy computational tasks to a secure, fortified cloud space. Early implementations will appear on Pixel 10 phones, enhancing tools such as Magic Cue and expanding language support for Recorder transcriptions. Google says the approach will enable richer, more personalized AI experiences without compromising privacy. Ler mais

A recent Deloitte survey shows that while a majority of U.S. consumers are actively using or experimenting with generative AI, they also express strong concerns about privacy, data security, and the trustworthiness of tech companies. More than half of respondents pay for AI services, yet many still verify AI‑generated information and are reluctant to share personal data. The findings highlight a paradox: rapid adoption of AI alongside growing skepticism about its impact and the motives of the firms behind it. Ler mais

Former Cold War shelters and abandoned mines are being transformed into high‑security data centers. Companies such as Cyberfort operate these subterranean facilities, offering protection against both cyber and physical threats. The hardened concrete walls, blast‑proof doors and strict access controls promise data survivability even in extreme scenarios. While the physical security is emphasized, the facilities also address regulatory concerns like data sovereignty and environmental impact by sourcing renewable energy and using closed‑loop cooling. The trend reflects growing anxieties over data loss and the need for resilient infrastructure. Ler mais

A recent Deloitte survey of U.S. consumers shows that while more than half are experimenting with or regularly using generative AI, a majority express concerns about rapid innovation, data privacy, and the accuracy of AI outputs. Around 40% of respondents pay for AI services, and many access the technology through mobile apps and websites. Trust remains fragile—privacy worries have risen, and users are reluctant to share sensitive personal data. Consumers indicate they are more likely to spend money with companies they trust, highlighting a tension between growing adoption and lingering skepticism. Ler mais

ChatGPT excels at answering questions and drafting text, but it falls short in critical areas such as diagnosing health issues, providing mental‑health support, handling emergency safety decisions, offering personalized financial advice, and processing confidential or regulated data. It also cannot replace legal professionals, nor should it be used for cheating in education, real‑time monitoring, gambling, or creating art that is passed off as original. Understanding these constraints helps users avoid costly mistakes and rely on qualified experts when needed. Ler mais

A potential deal to keep TikTok operating in the United States past the current deadline is moving forward. The arrangement would involve a consortium of U.S. firms, including Oracle, Silver Lake and Susquehanna International, taking control of the platform's American operations while ByteDance retains a minority board seat. The transition may be handled through an app update rather than a new download, but a new algorithm trained on fresh data is expected. Experts warn that technical hurdles and policy concerns could jeopardize the effort, leaving the platform's future in flux. Ler mais