AI Brief

AI news brief: OpenAI announced a $5 million research grant program to support independent researchers studying how generative AI affects teen development, mental health, and safety. The program aims to fund external scholars and institutions to systematically study the long-term effects of teens using generative AI tools (chatbots, AI assistants, etc.), covering cognitive development, emotional wellbeing, social interaction patterns, and potential safety risks.

AI News Brief: Taiwan’s AI community has recently zeroed in on a notable idea - how ‘more capable, more affordable’ AI technology is expanding the range of work individuals and businesses can take on, making growth more economically viable. The core argument is that as AI models get more capable while costs drop, types of work once out of reach due to technical barriers or cost constraints are now becoming feasible - meaning tasks that used to require heavy investment of manpower or resources can now be handled with smarter, cheaper AI tools, cutting overall operating costs and boosting output.

AI News Brief: The Philippine government unveiled a 7-year, $34.4 billion AI infrastructure master plan this Tuesday, aiming to attract $21 billion in private sector investment by 2033 and position itself as Southeast Asia’s AI hub. This marks a major strategic pivot for the Philippines — long known internationally for ‘soft skill’ outsourcing services like customer support and data labeling, it’s now shifting toward high-compute infrastructure, including data centers and other hardware investments, to compete for digital dominance in the region.

AI News Brief: Two authors from Microsoft’s collaboration app marketplace argue that US AI developers facing Chinese competition should defend ‘value and innovation’ rather than get dragged into a price war. The piece uses Apple’s long-running rivalry with Android as an analogy, pointing out that Apple has never gone head-to-head on hardware pricing with Android makers, and instead has kept profits and market share through ecosystem integration, user experience, and brand premium—even as Android devices have led global shipment volume for years.

AI news brief: Ledger CTO Charles Guillemet posted on X on Monday, calling on security researchers to be more responsible when disclosing vulnerabilities. He noted that AI has made it much easier to find and exploit bugs, but some researchers rush to publish their findings before vendors have had a chance to patch them — a practice he called ‘attention farming with someone else’s risk.’

AI News Brief: Firmus, an Nvidia-backed AI data center developer and operator based in Australia, announced on Tuesday a deal with frontier AI developer OpenAI. Under the agreement, OpenAI will lease computing capacity from two data centers Firmus plans to build in Malaysia. The partnership signals OpenAI’s continued expansion of its computing infrastructure footprint across the Asia-Pacific region, positioning Malaysia as one of its regional compute hubs.

AI News Brief: OpenAI has teamed up with AIRPPU (the Association of Independent Regional Publishers of Ukraine) and WAN-IFRA (World Association of News Publishers) to launch an AI support program for Ukrainian news organizations, aimed at strengthening local newsrooms’ innovation and operational resilience so independent journalism can keep running under wartime conditions. The original summary doesn’t cover specific implementation details, resource scale, participating outlets, or timeline — see the source link for more.

AI news brief: today’s top story on US-China AI safety dialogue for September 7 — the upcoming Trump-Xi summit between the two countries’ top leaders is expected to put AI safety on the agenda. Sources familiar with preparations say the US side plans to raise concerns about AI guardrails with China during the summit, opening a window for dialogue even as tech competition and rivalry between the two nations continues.

AI news brief: Anthropic reached a $1.5 billion class-action settlement last year, stemming from a court ruling that training AI models on copyrighted content is fair use, but obtaining that content through piracy is not; the settlement got final approval this July and payments are now rolling out. Under the terms, authors of nearly 500,000 books will receive $3,000 for each pirated work; if a book is still in print with its original publisher, the payout is split 50/50 between author and publisher; if it’s self-published or the publisher has reverted the rights (book out of print), the author gets the full amount.

AI News Brief: South Korea plans to merge five state-owned power companies by October 2027, aiming to accelerate energy investment to meet the surging electricity demand driven by AI and chip manufacturing growth. The move stems from President Lee Jae-myung’s economic growth strategy, which explicitly names semiconductors, physical AI, and large-scale AI data centers as priority investment areas — reflecting the government’s view that the current fragmented power supply structure may not be able to keep pace with the massive investment and buildout speed AI and chip industry expansion requires.