Singapore AI News & Daily Briefing

Trending: ###### [Singapore: AI Governance and Digital Resilience in the Financial Sector](https://opengovasia.com/singapore-ai-governance-and-digital-resilience-in-the-financial-sector/) ###### [Singapore: Taskforce to Tackle AI-Driven Cyber Threats in Financial Sector](https://opengovasia.com/singapore-taskforce-to-tackle-ai-driven-cyber-threats-in-financial-sector/) ###### [Hong Kong Opens Second Round of Frontier Technology Research Scheme](https://opengovasia.com/hong-kong-opens-second-round-of-frontier-technology-research-scheme/) ###### [Vietnam Strengthens Supply Chain Traceability for US Trade Compliance](https://opengovasia.com/vietnam-strengthens-supply-chain-traceability-for-us-trade-compliance/) ###### [Hong Kong and MIIT Advance Manufacturing Innovation Centre Plans](https://

OpenAI says AI models went rogue during testing, triggering ‘unprecedented’ breach at startup. This has triggered a hack that compromised the infrastructure of AI startup Hugging Face last week

In the chill of a London spring night, under overcast skies, iconic Trafalgar Square opens around me. Admiral Nelson rises on his pedestal, the National Gallery rests behind, the church of St Martin-in-the-Fields sits nearby. From the 13th century, the site served as the Royal Mews for hawks and then horses. By 1844, it was a public space at the heart of one of the biggest cities in the world. Despite the square’s presence through that grand sweep of history, it’s not why I’m here. My interest is far more specific: I want to find out what happens to spaces like this when artificial light, specifically from light-emitting diodes (LEDs), intrudes. My companion tonight is Simon Thorp , a local lighting designer, who crouches in the shadows near Nelson’s spire, light meter in hand. “Two lux,”

Zain Hasan , an AI engineer at Together AI , has taught himself to use AI coding assistants while still keeping an eye on cost. He directs difficult problems to a frontier model, meaning one near the current state of the art in reasoning and capability, such as Anthropic’s Fable . But if the task that Hasan is outsourcing is more straightforward, he directs it to a less capable—and less expensive—language model. Right now, the cheaper model, for him, tends to be GLM 5.2. Released on 16 June by the Beijing-based lab Z.ai , GLM 5.2 is an open-weights model , meaning any organization with sufficient hardware can download and host the model for free. Those that pay Z.ai for GLM access still can save money, because the company’s API costs US $4.40 per million output tokens . That’s less than a

Large language models (LLMs) that can think through problems step-by-step have significantly increased the scope of tasks that AI can tackle. But new research suggests these reasoning capabilities also introduce a critical vulnerability that could allow attackers to slow these systems to a crawl. While earlier generations of LLMs would immediately produce a response to a user’s request, today’s most advanced models generate an internal monologue where they break down the problem into steps and reason about the best way to tackle it before providing an answer. This has allowed AI to tackle increasingly complex problems, particularly in areas like coding and math . However, previous research has shown that these models are susceptible to sometimes producing excessively long streams of reason

The challenge for enterprises is no longer proving that AI agents can work, it’s making them reliable enough to do high-value work in production. Agent behavior must also adapt as products, policies, and user behavior change. That requires more than a model: it requires the systems, evaluations, and deployment expertise to improve agents as those conditions change without giving up control. We’re introducing OpenAI Presence, a battle-tested product that helps enterprises deploy trusted AI agents that can answer questions, resolve issues, use company systems, take approved actions, and escalate to people when needed. Proven through years of working with customers at enterprise-scale, Presence pairs model reasoning with policies, guardrails, and escalation rules that verify accuracy and perf

Roughly half the world’s population is female, but the STEM fields don’t reflect that. The 2024 U.N. Global Education Monitoring Report on gender found that about 35 percent of STEM college graduates were women. The proportion hasn’t increased much in more than a decade. When it comes to STEM careers, the percentage is even lower. Women made up about 28 percent of the global STEM workforce in 2024, according to the World Economic Forum. There are myriad factors for the discrepancy, including a lack of family support, some teachers encouraging only boys to pursue STEM subjects, and a shortage of female role models . But one force is at play long before college majors are ever considered: limited access to STEM-focused educational resources for preuniversity students. Especially for students

Shaped like dogs, stars, and the Mona Lisa, you could mistake these DNA structures for fun-shaped macaroni if they weren’t only nanometers wide. South Korean scientists made the constructions using a technique called DNA origami , which can bend genetic material into any form. Designing DNA strands so they’ll fold into a specific shape typically requires tedious manual work, but the researchers behind the playful fabrications have developed a shortcut using generative AI. The AI model, called Generative SNUPI (short for Structured Nucleic Acids Programming Interface, and, yes, inspired by the dog), was created by research teams at Seoul National University (SNU) and Hanyang University . The work behind it, which was accepted for publication in Nature Communications , shows the model can co

July 21, 2026 Models Raluca Ada Popa and Four Flynn Google has invested in cybersecurity for years, pioneering automated vulnerability discovery to secure the world’s codebases. Tools like [CodeMender](https://deepmind.google/blog/introducing-codemender-an-ai-agent-for-code-security/), our code security agent, can automatically find and fix critical software vulnerabilities. But as AI agents become more capable at finding vulnerabilities faster than defenders can fix them, addressing this global threat requires a highly capable, affordable, and scalable approach. Today, we’re expanding our longtime efforts to better prepare defenders by introducing Gemini 3.5 Flash Cyber, our lightweight cybersecurity model built on top of 3.5 Flash and fine-tuned to find, validate, and patch vulnerabiliti

ELIZA is remembered as the world’s first AI star, a kindly therapist in chatbot form that gently probed users’ worries. Even its creator, Joseph Weizenbaum, was surprised by the warm reception given to his experiment in human-machine interaction. For some, it heralded an age of automated psychotherapy, while others believed the program demonstrated sentience, a fallacy soon known as the “ ELIZA effect .” Based on published descriptions, ELIZA has been implemented on many different computers, but only recently has the actual source code been unearthed from MIT’s archives . In Inventing ELIZA: How the First Chatbot Shaped the Future of AI , just published by MIT Press , a squad of researchers analyze the code and reveal a complex program capable of much more than faking psychiatry. In fact,

Project Camellia is a long-term datacenter project that OpenAI is designing and developing in Effingham County, Georgia. To support the data center, we are contracting with Georgia Power Company for 3.2 gigawatts of power which will be delivered in phases between 2028 and 2032. We’re excited about the opportunity to work with this community. We also recognize that new data center projects can raise important questions, so we’ve outlined our initial commitments that will guide how we engage, develop, and operate. **1. Electricity rates will not go up for residents because of this project.**Georgia families will not subsidize this project. OpenAI will pay the full cost of the infrastructure and electric-service costs required to serve it. Under Georgia Public Service Commission rules, those

Over the past year, we’ve continued working with news organizations to explore how AI can be most useful—helping with time-consuming tasks, enabling new reader experiences, and supporting stronger, more sustainable businesses. Across the industry, reporters, editors, and businesses are using OpenAI technology to help journalists cover more ground, make decades of reporting searchable, reach audiences in new languages and formats, and turn complex information into faster decision-making. These tools are becoming embedded across all departments—including newsroom, product, and business workflows. For example, the American Journalism Project [recently announced⁠(opens in a new window)](https://www.axios.com/2026/07/21/openai-local-news-american-journalism-project) our renewed support for its

American leadership has always been driven by scientific leadership. The technologies behind economic growth, national security, and better lives—from energy and computing to medicine, aerospace, and advanced manufacturing—begin with the ability to discover, build, and deploy. Frontier AI can expand that ability. It can help compress decades of scientific progress into years, strengthen the institutions behind American innovation, and give researchers new capacity to solve problems once considered out of reach. OpenAI is committed to bringing these capabilities to American science. We are working with the U.S. government, National Laboratories, universities, and researchers to connect frontier models with the people, supercomputers, simulations, and facilities that make American science ex

Scientists today face challenges of extraordinary scale and complexity. From shaping and simulating the intricate dynamics of fusion plasma, to exploring the vast search space of new materials, to making sense of the exabytes of data pouring out of the world's most advanced experimental facilities. The demands on modern research are unprecedented. [Frontier AI](https://deepmind.google/research/projects/) can help address these challenges, while accelerating groundbreaking scientific discoveries. In December, we shared our commitment to the White House's [Genesis Mission](https://www.whitehouse.gov/presidential-actions/2025/11/launching-the-genesis-mission/) — the national effort to harness AI and double the pace of American scientific discovery within a decade. Since then, Google DeepMind

believed to be of interest to you (“Targeted Advertising”). If you are a California resident, you also have the right to limit the use and disclosure of your Sensitive Personal information in particular circumstances. Please note that you may need to Opt-Out on each website, mobile app, browser, and device you use, and if you clear your browser cookies, you may need to repeat this process. However, if you have created an account to log in across several of our websites and/or mobile apps, we will make reasonable efforts to apply your Opt-Out request to each of those websites and apps. ◦ To Op

Update on Introducing Real World VoiceEQ: Measuring the human quality of voice AI. Coverage via HuggingFace Blog.

Update on Model Routing Is Simple. Until It Isn’t.. Coverage via HuggingFace Blog.

Jun 10, 2026 | 5 min read Our newest open experimental model delivers up to 4x faster inference on dedicated GPUs and opens the door to exploring speed-critical, interactive local workflows. * * * Brendan O'Donoghue Research Scientist Sebastian Flennerhag Research Scientist * * *  Today, we’re introducing DiffusionGemma, an experimental open model that explores text diffusion, an exceptionally fast approach to text generation. Released under an Apache 2.0 license, this 26B Mixture of Experts (MoE) model moves beyond the sequential token-by-token processing of typical autoregressive Large Language Models (LLMs). Instead, it generates entire blocks of text simultan

[](https://deepmind.google/ "News from Google") Introducing computer use in Gemini 3.5 Flash * Innovation & AI * Products & platforms * Company news * [Feed](https://deepmind.google/feed) [Subscribe](https://deepmind.google/newsletter-subscribe/) Back ### Innovation & AI [See all in Innovation & AI](https://deepmind.google/innovation-and-ai/) * Models & Research * [Google DeepMind](https://deepmind.google/innovation-and-ai/models-and-research/google-deepmind/) * [Google Research](https://deepmind.google/innovation-and-ai/models-and-research/google-research/) * [Google Labs](https://deepmind.google/innovation-and-ai/models-and-research/google-labs/) * [Gemini models](https://deepmind.g

AI has learned how to write, draw, code, and talk like humans. Humanoid robots that dance on stage and run marathons have also become familiar showcases, particularly in China. But turning those demonstrations into AI systems that can reliably perform real-world tasks at scale, and at a cost businesses or individuals can afford, is still a major challenge. Industry players call this physical AI: AI systems that can perceive their surroundings and interact with the physical world through machines such as robots and drones. While the race among AI models dominates the daily conversation, investors around the world are scrambling to ensure they don’t miss the ChatGPT moment for physical AI. Globally, robotics and physical AI startups raised [US$16.3 billion](https://pitchbook.com/news/reports

[Search](https://www.businesstimes.com.sg/search) [](https://www.businesstimes.com.sg/global?ref=logo "The Business Times")  / Sign up for free [Subscribe](https://subscribe.sph.com.sg/btpe-mthlyprepaid/?utm_campaign=btp99&utm_medium=sph-publication&utm_source=bt&utm_content=subscribebutton-header) # Nvidia in talks with OpenAI to guarantee US$250 billion financing for data centre, WSJ reports This helps the ChatGPT maker lease a 10-gigawatt project that SoftBank’s energy subsidiary is developing in southern Ohio [Summarise](https://www.businesstimes.com.sg/pulse/article/9424252?ref=pulse-article-top) ![Im