Data privacy

Apple

2026-09-10 03:06:06

Apple CEO John Ternus says iPhone is the best AI device available today

Apple CEO John Ternus said at the opening of the company’s Surprise and Shine event that the iPhone is already the best AI device available today, while arguing that Apple places more weight on user privacy than its rivals. His remarks came after he succeeded Tim Cook as Apple’s chief executive, with Cook now serving as executive chairman. Ternus said Apple does not plan to build entirely new hardware for AI. Instead, the company wants the iPhone to remain the main way consumers interact with the technology. He described the ideal AI product as an “intelligent personal hub” that understands a user’s personal context, stays with them at all times, runs AI models on-device, connects to cloud-based models, and includes a large display, high-quality cameras and microphones, long battery life, ease of use, and close interoperability with other devices, apps, and services. He added that no product is better suited than the iPhone for that role, and said Apple Intelligence is designed to run on-device whenever possible because trust weakens once user data is no longer under the user’s control.

820

KYC

2026-09-09 13:02:49

Coin Center’s Laz Pieper calls for a new approach to KYC data collection

KYC data has become a highly attractive target for hackers, according to a CoinDesk opinion piece centered on a proposal from Coin Center’s Laz Pieper. He argues that the way identity information is collected should change, with the focus shifting away from gathering large pools of underlying personal data. Instead, privacy-preserving identity verification systems could let individuals prove only the specific facts a service needs to know. Under that model, users would keep control of the underlying information rather than handing it over in full. The argument is that identity checks do not always require broad disclosure of sensitive records, and systems can be designed to verify limited claims while reducing the exposure created by centralized data collection. The piece frames current KYC repositories as an irresistible honeypot for attackers and presents privacy-preserving verification as an alternative that could reduce that risk while still meeting service-level identity requirements.

250

AI agents

2026-09-04 00:11:08

Legal analysis weighs whether AI agents need both user consent and platform permission to act on a user’s behalf

A legal analysis published by Internet Law Review examines a question that is moving closer to real-world deployment as AI agents begin handling tasks for users inside third-party apps and websites: is explicit user authorization enough, or must the operator of the app or website also give permission? The article, republished by MarsBit, breaks the issue into two technical paths. Under an API-based model, the platform must open interfaces, define access rights, and authenticate the service provider, which means the agent service can only function with both the user’s authorization and the platform’s cooperation. In that setting, “dual consent” exists as a practical reality. Under a GUI-based model, by contrast, the agent runs through the user’s own device, reads on-screen content through screenshots, OCR, and interface recognition, and simulates clicks and typing without its server directly connecting to the third-party platform. The piece argues that this makes the agent closer to a tool used by the customer than an independent actor or legal proxy. It also reviews platform terms that ban or restrict AI agents, saying their enforceability should be judged under China’s Civil Code rules on standard-form clauses, with close attention to notice obligations, scope, and whether the clause unreasonably strips users of core rights.

1100

Policy Regula

2026-09-02 08:29:25

Lawyer Lin Shanglun says AI moats should be built around offline scarcity, trust, and compliance

Lawyer Lin Shanglun argues that as generative technology drives the cost of producing text, code, and charts close to zero, the assets and services with the strongest defenses may be the ones algorithms cannot easily replicate online. In his article, he points to the continued strength in physical Pokemon card and vintage sports card auctions as a case study in scarcity, shared historical value, and offline authentication. Lin says the same logic applies to professional services and enterprise software. In his view, the most durable barriers come from three areas: strict data privacy and international cybersecurity compliance, face-to-face trust built through offline business relationships, and non-public practical knowledge accumulated by experienced professionals. Public information can be read by models, he writes, but the details that often matter most in legal, medical, and operational work are not available on the open internet. He goes on to argue that AI builders should design products in reverse, embedding these “AI-resistant” elements into the system itself. That includes expert verification workflows, private feedback loops that turn offline know-how into proprietary assets, contract-level accountability structures, and human-in-the-loop delivery models that combine automation with expert judgment.

980

Anthropic

2026-09-01 18:23:47

Anthropic Revises Data Retention Policy in Response to Customer Feedback

AI company Anthropic has revised its data retention policy, directly responding to feedback from enterprise clients. The update aims to enhance trust and transparency in data handling. The Enterprise Frontier Safeguards changes were shaped by customer input. The new policy clearly defines rules for storing, using, and deleting user data. The news was reported by CNBC.

770

John Deere

2026-09-01 16:03:08

John Deere Tests AI Assistant JD to Help Farmers Boost Yields

John Deere, the agricultural machinery manufacturer, is testing an AI assistant named "JD" aimed at helping farmers improve profitability. The assistant uses users' farm, equipment, and operational data to answer questions about best practices and historical trends, covering equipment settings, fuel usage, and harvest timing. The company has launched an early access program for the chatbot and outlined a ten-point "Farmer Data Promise," vowing not to sell user data and allowing farmers to control their data. The specific AI technology behind the platform has not been disclosed.

770

UC Berkeley

2026-08-23 10:55:51

UC Berkeley and UT Austin researchers release edge-native MoE inference engine FreeToken

Researchers from the University of California, Berkeley and the University of Texas at Austin have introduced FreeToken, an edge-native mixture-of-experts inference engine designed to turn personal computers into a unified elastic inference platform. According to the release cited by Techub, the system can run the 753B-parameter GLM-5.2 model on a single workstation GPU, a 284B model on a gaming desktop, and a 35B model at interactive speed on a laptop GPU with 8GB of VRAM. FreeToken has been open-sourced under the Apache-2.0 license on GitHub and published on PyPI. The team also provides one-click desktop applications for Windows and Linux. Its command-line interface supports Linux x86_64 systems and NVIDIA GPUs, and the ft serve command can expose an API endpoint on port 1919 that is compatible with OpenAI and Anthropic. The project is aimed at individual developers, startups, and engineering teams at small and medium-sized businesses, with a focus on privacy-sensitive use cases such as healthcare, legal work, defense, finance, and intellectual-property-heavy R&D. Example applications include local coding agents, private code review, offline contract analysis, and synthetic data generation.

1340

Google

2026-08-18 17:19:44

Google Plans $10 Million Purchase of Spirit Airlines Data for AI Training

Google plans to pay $10 million for business data from Spirit Airlines, adding another example of major technology companies buying large data sets to improve artificial intelligence systems. According to CNN, the proposed deal covers internal communications and company records, including emails, spreadsheets, bookings, frequent-flyer information, and employee human resources data. Reuters separately reported that the package also includes Microsoft Teams messages and calendars, along with marketing, productivity, and operations records. Google told CNN that it acquired part of an enterprise data set from Spirit Airlines because the material could help improve its products and AI models. The company said it will not receive personal information, and the data is expected to be stripped of details that could identify individuals before transfer. The sale has not closed yet and still needs court approval. U.S. Bankruptcy Judge Sean Lane is scheduled to review the transaction at a hearing on Wednesday. Spirit halted operations in May and has been selling remaining assets through bankruptcy. Reuters said AI data company Mercor also competed for the assets with a reported $7.5 million bid. The agreement fits into a broader pattern. Google struck a reported $60 million-a-year deal with Reddit in February 2024, OpenAI reached its own Reddit agreement in May 2024, and the Wikimedia Foundation announced content access agreements in January with Microsoft, Google, Amazon, Meta, and others.

500