Artificial intelligence

Malaysia under scrutiny over semiconductor exports to China

Malaysia plans to tighten regulations on semiconductor shipments following US concerns over the potential transfer of high-end Nvidia chips to China.

Trade Minister Zafrul Aziz stated that the United States has urged Malaysia to closely monitor shipments, ensuring that advanced AI chips do not end up in unauthorised locations.

The move comes amid increasing global scrutiny over AI-related technology exports.

Authorities in Malaysia are also investigating whether local laws were breached in a case involving servers linked to a Singapore fraud investigation.

The case involves transactions worth $390 million, and reports suggest that some servers may have contained Nvidia chips subject to US export controls. Singapore media have linked the matter to potential transfers to Chinese AI company DeepSeek.

The United States has been tightening restrictions on advanced semiconductor exports to China, particularly chips crucial to AI development.

Malaysia’s role as a key semiconductor hub has drawn greater attention, with US officials pushing for stricter oversight.

The government is expected to introduce measures to ensure compliance with international regulations while maintaining its position in the global chip supply chain.

For more information on these topics, visit diplomacy.edu.

AI agents take centre stage in Oracle fusion

Oracle has launched its AI Agent Studio, a new platform designed to let businesses orchestrate and customise AI agents within its Fusion Applications suite.

Announced during the OracleCloud World Tour in London, the studio enables companies to coordinate teams of AI agents that handle tasks across enterprise resource planning, HR, supply chain, and customer experience systems.

The AI Agent Studio allows businesses to adapt prebuilt Oracle agents to suit their own processes. Users can modify agents by adjusting logic, integrating external tools, or adding custom prompts.

It also offers flexibility in choosing from a range of large language models optimised for Oracle or industry-specific use cases, such as Llama and Cohere.

Oracle’s move builds on earlier AI deployments in its cloud applications, where agents have been embedded to manage routine operations like invoice processing or recruitment steps.

The new platform advances that effort by allowing these agents to operate collaboratively and be tailored to more complex workflows.

Industry leaders including Accenture, Deloitte, and PwC have praised the development, calling it a significant step toward smarter enterprise automation.

Analysts echo this sentiment, noting that Oracle’s approach allows businesses to maximise AI efficiency across departments without added cost, offering a powerful edge in today’s rapidly evolving digital workplace.

For more information on these topics, visit diplomacy.edu.

FuriosaAI rejects $800m acquisition offer from Meta

FuriosaAI, a South Korean startup specialising in AI chips, has reportedly turned down an $800 million acquisition offer from Meta.

Instead of selling, FuriosaAI plans to continue developing its AI chips. Disagreements over post-acquisition business strategy and organisational structure were reportedly the cause of the breakdown in negotiations, rather than issues over price.

Meta, which has been trying to reduce its reliance on Nvidia for chips specialised in training large language models (LLMs), unveiled its custom AI chips last year. The company also announced plans to invest up to $65 billion this year to support its AI initiatives.

FuriosaAI, founded in 2017 by June Paik, who previously worked at Samsung Electronics and AMD, has developed two AI chips—Warboy and Renegade (RNGD).

The startup is also in talks to raise approximately $48 million and is planning to launch the RNGD chips later this year, with LG AI Research already testing them for use in its AI infrastructure.

FuriosaAI’s decision to focus on expanding its chip production signals its confidence in competing with giants like Nvidia and AMD in the rapidly growing AI hardware market.

For more information on these topics, visit diplomacy.edu.

Apple plans to add cameras to future Apple Watch

Apple is reportedly planning to introduce cameras to its Apple Watch lineup within the next two years, integrating advanced AI-powered features like Visual Intelligence.

According to Bloomberg’s Mark Gurman, the standard Apple Watch Series will have a camera embedded within the display, while the Apple Watch Ultra will feature one on the side near the digital crown.

These cameras will allow the smartwatch to observe its surroundings and use AI to provide real-time, useful information to users.

Apple is also exploring similar camera technology for future AirPods, aiming to enhance their functionality with AI-driven capabilities.

The concept builds on the Visual Intelligence feature introduced with the iPhone 16, which allows users to extract details from flyers, identify locations, and more using the phone’s camera.

While the current system relies on external AI models, Apple is working on its in-house AI technology, and it is expected to power these features by 2027, when the camera-equipped Apple Watch and AirPods are likely to be released.

The move is part of Apple’s broader push into AI, led by Mike Rockwell, who previously spearheaded the Vision Pro project.

Rockwell is now overseeing the upgrade of Siri’s language model, which has faced delays, and contributing to visionOS, the operating system expected to support AI-enhanced AR glasses in the future. Apple’s increasing focus on AI suggests a shift towards more intelligent, context-aware wearable devices.

For more information on these topics, visit diplomacy.edu.

Google expands Gemini with real-time AI features

Google has begun rolling out real-time AI features for its Gemini system, allowing it to analyse smartphone screens and camera feeds instantly. These capabilities, which will be available to select Google One AI Premium subscribers, build on the company’s earlier ‘Project Astra’ demonstration.

The live video feature will enable Gemini to interpret smartphone camera feeds in real time, providing users with instant answers and insights.

The new functionality also allows users to engage in back-and-forth conversations with Gemini based on their screen’s content. A Reddit user recently demonstrated the ‘Share screen with Live’ feature, accessible via the Gemini overlay, showcasing its ability to process and respond to information directly from a device’s display.

Google has confirmed that these updates will first roll out to Gemini Advanced subscribers under the Google One AI Premium Plan, with Pixel and Galaxy S25 owners among the first to gain access.

In addition to real-time AI video capabilities, Google has introduced ‘Canvas,’ a tool designed to help users refine documents and code seamlessly. Canvas allows for real-time edits and streamlines the process of developing prototypes for web apps, Python scripts, and other digital projects.

Another notable addition is ‘Audio Overview,’ which transforms written documents, slides, and research reports into podcast-style discussions between two AI-generated hosts.

An innovation like this aims to make complex information more engaging and accessible by delivering content in a conversational format. Google continues to expand Gemini’s capabilities, reinforcing its position at the forefront of AI-driven user experiences.

For more information on these topics, visit diplomacy.edu.

New Airbyte connectors support AI and data privacy

San Francisco-based data startup Airbyte has unveiled a new set of enterprise tools aimed at helping companies move and manage data more securely, especially as AI becomes more central to operations. The updates, announced Thursday, include new connectors for apps such as NetSuite, SAP, and ServiceNow, as well as support for extracting unstructured data from platforms like Google Drive and SharePoint.

A key highlight of the release is compatibility with Apache Iceberg, an open-source format that enables businesses to centralise data into a single, AI-compatible “lakehouse.” This allows companies to better control how and where their data flows while preserving the flexibility needed for high-performance analytics and machine learning.

Airbyte co-founder and CEO Michel Tricot stressed the importance of data sovereignty in an AI-driven era. He noted that while AI tools can be powerful, giving away sensitive internal data, like employee compensation or strategic business metrics, to external services is a risk many companies are no longer willing to take. Airbyte’s approach ensures that only the enterprise sees and manages its data pipelines.

Founded in 2020, Airbyte now serves over 7,000 enterprise clients, including names like Invesco and Calendly, and has secured more than $181 million in funding. As businesses continue to prioritise secure, scalable infrastructure for AI, Airbyte’s offerings are positioning it as a go-to partner for data portability without compromise.

For more information on these topics, visit diplomacy.edu.

Web access comes to Anthropic Claude chatbot

Anthropic has added web search functionality to its Claude chatbot, marking a major step forward for the AI assistant. The feature is now available in preview for paying users in the US, with plans to expand to free users and more countries in the near future. Once enabled via profile settings, Claude will automatically pull information from across the web to help respond to queries, complete with source citations.

The capability is limited to the newest version of Claude, the Claude 3.7 Sonnet model. According to Anthropic, the goal is to combine Claude’s conversational skills with up-to-date information, allowing it to deliver more accurate and relevant AI responses. Early testing suggests the web search doesn’t always activate, particularly for current events, but when it does, it sources from a range of platforms including social media and news outlets.

This move brings Claude in line with competing chatbots such as ChatGPT, Gemini and Le Chat, all of which already integrate web access. Anthropic had previously resisted the feature, insisting that Claude was meant to be self-contained, but market pressure appears to have shifted that stance.

Despite the new functionality, concerns persist over accuracy. Studies have shown that leading chatbots regularly produce incorrect or misleading responses when sourcing from the internet. Anthropic acknowledges these risks but believes that providing clear citations and structured output will help mitigate confusion and improve user trust.

For more information on these topics, visit diplomacy.edu.

Apple accused of misleading AI advertising

Apple is facing a class-action lawsuit in the United States over delays in delivering its much-promoted Apple Intelligence features.

The legal action, filed in a US based San Jose federal court, claims the company misled customers by advertising advanced AI tools that have yet to materialise on supported devices.

The complaint argues that buyers of new iPhones and other Apple products were promised ‘transformative’ AI capabilities at launch, only to find these features were either severely limited or completely absent.

According to the plaintiffs, Apple’s marketing created a “reasonable consumer expectation” that was ultimately not met.

This legal challenge adds to mounting pressure on the company, which has struggled to roll out its next-generation AI tools.

A recent Bloomberg report suggested internal tensions, revealing that CEO Tim Cook has reportedly lost confidence in AI chief John Giannandrea’s ability to deliver on the company’s ambitions.

The case reflects growing scrutiny of tech firms’ promises around AI, especially as consumer trust becomes more closely tied to the reality behind flashy announcements.

For more information on these topics, visit diplomacy.edu.

Chinese AI startup Manus gains recognition in state media

Chinese AI startup Manus has registered its AI assistant for the domestic market and gained state media attention, signalling government support for its expansion.

The company, which recently gained global recognition, was featured on China’s CCTV in a segment comparing its AI agent to the chatbot made by DeepSeek’s chatbot.

Beijing’s growing interest in homegrown AI firms follows the success of DeepSeek, which developed competitive AI models at a lower cost than its United States counterparts.

Manus has made headlines for developing an AI agent capable of making autonomous decisions with minimal prompting, a step beyond traditional chatbots like ChatGPT.

The firm recently announced a partnership with Alibaba’s Qwen AI team, a move that could accelerate its rollout in China. Currently, the AI agent is available only through invite codes, with a waiting list of two million users.

Regulatory approval was also secured for Monica, Manus’ AI assistant, allowing its launch in China. All generative AI applications in the country must comply with strict content regulations to align with Beijing’s policies.

Manus’ growing presence in China highlights the government’s strategy of fostering domestic AI innovations while ensuring regulatory oversight.

For more information on these topics, visit diplomacy.edu.

Google adds Mind Maps to NotebookLM

Google has unveiled a new feature called Mind Maps for its AI-powered research tool, NotebookLM. Mind maps are visual diagrams that help users understand complex subjects by displaying ideas and their connections.

An addition like this follows the recent release of Audio Overviews, which provide AI-generated podcasts summarising key points from documents, articles, and videos.

NotebookLM, which works in both free and paid versions, assists users in summarising content and offering interactive conversations with AI to deepen understanding.

The new Mind Maps feature lets users generate and explore visual connections between ideas. Once created, users can zoom, expand or collapse branches, and click on nodes for detailed information on specific topics.

The feature is particularly useful for students or anyone who needs to absorb a lot of information quickly. With the combined power of Mind Maps and Audio Overviews, NotebookLM offers a multi-faceted approach to learning, making it easier to navigate and retain key insights.

For more information on these topics, visit diplomacy.edu.