Morning Brief

1. OpenAI publishes an article exploring policy recommendations for the AI era
2. OPPO launches OPPO A6k
3. China’s Ministry of Industry and Information Technology issues a risk alert on specific iOS versions regarding vulnerability exploitation
4. LinkedIn accused of scanning users’ browser extensions
5. Cyberspace Administration of China proposes stronger regulation of digital virtual human services
6. Ten government departments issue measures on AI technology ethics review and services
7. Two updates from Anthropic
8. News Worth a Quick Look

---

OpenAI publishes an article exploring policy recommendations for the AI era

On April 6, OpenAI released an article titled Industrial Policy for the Intelligence Age. The article argues that as humanity moves toward superintelligence, incremental policy updates are no longer sufficient. OpenAI proposes a set of human-centered policy recommendations aimed at expanding opportunity, sharing prosperity, and building resilient institutions to ensure that advanced AI benefits everyone. The vision includes broadly shared prosperity and improved quality of life, mitigating risks through new institutions, technical measures, and governance frameworks, and democratizing access to and control over AI. The article also suggests concrete measures such as establishing a “public wealth fund,” implementing automation taxes, and translating efficiency gains into long-term worker benefits, such as a four-day workweek. Source

---

OPPO launches OPPO A6k

On April 6, OPPO launched the OPPO A6k smartphone. The device is powered by the MediaTek Dimensity 6300 processor and comes in three configurations: 8GB+256GB, 8GB+512GB, and 12GB+256GB. It is available in Twilight Blue, Seashell White, and Dawn Gold. The phone features a 6.75-inch HD+ 120Hz display, a 7000mAh battery with 45W fast charging, and supports IP69, IP68, and IP66 water resistance. It also supports microSD expansion up to 2TB and allows touch input with gloves (under 5mm thickness) or wet hands. Pricing starts at 1,999 RMB. Source

---

China’s Ministry of Industry and Information Technology issues a risk alert on specific iOS versions regarding vulnerability exploitation

On April 3, China’s Ministry of Industry and Information Technology (MIIT) vulnerability platform (NVDB) detected active exploitation of vulnerabilities targeting Apple devices. These attacks could result in data theft and full system compromise. Affected devices include iPhones and iPads running iOS 13.0 through 17.2.1. Attackers may use SMS, email, or malicious web pages to lure users into opening links in Safari, leveraging multiple vulnerabilities to install remote access trojans, steal sensitive information, and gain full control of the device. Users are advised to update to the latest security version promptly. Source

---

LinkedIn accused of scanning users’ browser extensions

According to a report by Bleeping Computer on April 3, Fairlinked e.V., a self-described European association of LinkedIn business users, released a report titled “BrowserGate,” accusing LinkedIn of injecting JavaScript into user sessions to detect thousands of browser extensions and link the results to identifiable user profiles. The report claims LinkedIn scans over 200 competing tools and, due to the platform’s linkage with professional identity data, may have accessed customer lists of numerous software companies without users’ knowledge. Independent testing by Bleeping Computer found that JavaScript files were indeed checking up to 6,236 browser extensions. Source

---

Cyberspace Administration of China proposes stronger regulation of digital virtual human services

According to Xinhua News Agency, on April 3, the Cyberspace Administration of China released a draft of the Administrative Measures for Digital Virtual Human Information Services (for public consultation). The draft stipulates that no organization or individual may use digital virtual human services to infringe on others’ personality rights through defamation or distortion. Without consent, services must not create digital avatars that can identify specific individuals. It also requires explicit and informed consent, with clear explanations of purpose, necessity, and impact on personal rights. Upon withdrawal of consent, service providers must delete related personal data and cease all use unless otherwise required by law. In addition, unless otherwise agreed, the digital avatar should be deactivated. Source

---

Ten government departments issue measures on AI technology ethics review and services

According to People’s Daily on April 4, the Ministry of Industry and Information Technology and nine other departments have jointly issued the Measures for the Ethical Review and Services of Artificial Intelligence Science and Technology (Trial). The measures state that AI ethics reviews should focus on six key aspects: human well-being, fairness and justice, controllability and trustworthiness, transparency and explainability, accountability and traceability, and privacy protection. These include evaluating whether AI activities have scientific and social value; whether the selection of training data and the design of algorithms, models, and systems are reasonable; whether the intended use and operational logic of algorithms and systems are properly disclosed; and whether sufficient measures are in place to ensure effective protection of personal data. Source

---

Two updates from Anthropic

On April 4, Anthropic executive Boris Cherny announced on X that starting from 12:00 PM PT on April 5, due to surging usage, Claude subscription plans will no longer include usage generated by third-party tools such as OpenClaw. Users will need to purchase additional usage packages or pay via Claude API keys. Cherny stated that Claude subscriptions were not designed for third-party tool usage patterns and that priority will be given to users of their core products and APIs. Source

On April 6, Anthropic released a press statement announcing a multi-gigawatt agreement with Google and Broadcom for next-generation TPU compute capacity expected to come online in 2027. Anthropic had already partnered with Google Cloud for TPU capacity in October last year. The statement also revealed that demand for Claude surged in 2026, with projected annualized revenue exceeding $30 billion and more than 1,000 enterprise customers spending over $1 million annually. By comparison, annualized revenue at the end of 2025 was around $9 billion. Source

---

News Worth a Quick Look

• According to CLS, Foxconn has begun trial production of a foldable iPhone. Source
• The Linux 7.1 kernel will remove support for Intel 486 CPUs. Source
• Linus Torvalds confirmed that the Linux 7.0 kernel will officially release its final version next week. Source
• According to The Information, influenced by “vibe coding,” App Store submissions increased by 84% year-over-year, with approximately 600,000 submissions in 2025—up 30% compared to 2024. Source
• Google has added a search feature to Play Store app reviews, allowing users to quickly find reviews containing specific keywords. Source
• On April 5, Samsung announced that the Samsung Messages app will be discontinued in July 2026. Future Galaxy devices will replace it with Google Messages. Due to compatibility limitations, Tizen OS watches released before Galaxy Watch4 will no longer be able to view full message histories, though SMS sending and reading will remain available. Source
• LinkedIn user Tey Bannerman noted that Microsoft has used the name “Copilot” across 78 different products, a number later updated to 80 through comments. Source
• Netflix will launch a global game hub app called “Playground” on April 28, designed for children under eight. Available to all Netflix subscribers, it will initially feature games based on Peppa Pig, Sesame Street, and Dr. Seuss. Source
• On April 6, leaker Ice Universe claimed that the Samsung Galaxy S27 series will introduce a new Pro model, essentially an Ultra variant without the S Pen. The privacy display feature introduced in the S26 Ultra may also be included. Source