NewsFeed

April 17, 2025

Google's AI Promotes Anti-Abortion Views, Report Finds

Google's AI technology has been found to be echoing anti-abortion messaging, according to a report. The report highlights how Google's AI-powered chatbots and search results have inadvertently promoted pro-life views on abortion, sparking concerns over the company's role in shaping public opinion. The issue was discovered by researchers who analyzed Google's AI responses to queries related to abortion and found that they often included information that supports anti-abortion views.

https://www.livemint.com/technology/tech-news/googles-ai-overviews-echo-anti-abortion-messaging-report-finds-11744905429334.html

April 17, 2025

Iron's 'EagleEye' Robot Revolutionizes Industry Standards

China has developed a humanoid robot with supercomputer-like brainpower and eagle-eye vision, surpassing current technology. The robot was designed by Iron, a Chinese company, and is capable of performing complex tasks. This breakthrough could have significant implications for various industries such as healthcare, manufacturing, and space exploration.

https://interestingengineering.com/military/stark-virtus-kamikaze-drone-germany

April 17, 2025

OpenAI Unveils AI Models o3 and o4-mini

OpenAI has launched two new AI models, o3 and o4-mini, designed for coding, math, and other complex reasoning tasks. The company claims these models are among its most advanced yet, with o3 being highlighted as the top-performing reasoning model to date. o4-mini is a more affordable version that still delivers impressive results in similar areas, following just two days after introducing GPT-4.1.

https://indiatoday.in/technology/news/story/openai-rolls-out-new-o3-and-o4-mini-ai-models-for-coding-math-and-other-tasks-2710262-2025-04-17

April 17, 2025

OWASP Gen AI Security Project Unites Industry Leaders

The OWASP Gen AI Security Project is leading efforts to build secure approaches to Generative AI through open source and community collaboration. The project brings together leaders from organizations such as Google, Cisco, Intel, Kainos, and others to discuss collaborative security strategies for LLM and Gen AI. A half-day summit at the RSA Conference will explore best practices in LLM and Gen AI security, while an interactive workshop invites community members to explore real-world threats tied to agentic AI systems. The project aims to provide security professionals with tools and insights to secure next-generation AI-powered applications through open education and expert-driven collaboration.

https://www.prnewswire.com/news-releases/owasp-gen-ai-security-project-announces-nine-new-sponsors-and-major-rsa-conference-presence-to-advance-generative-ai-security-302431171.html

April 17, 2025

Trump Administration Fails to Retain AI Talent

The Trump administration can sustain America's advantage in the AI race by fixing immigration to retain top talent, investing in STEM education and AI literacy, and democratizing access to high-performance computing. The US needs an immigration system that caters to high-skilled individuals, with bipartisan support for legislation to keep international students. Education is key, with only 30% of students meeting science proficiency standards, and the country needs national STEM mobilization to improve public school pipelines and expand access to computer science courses. Democratizing access to high-performance computing is also crucial, as most student researchers and public universities lack access to the compute power available to large technology firms and government researchers.

https://www.csis.org/analysis/winning-tech-race-china-requires-more-restrictions

April 17, 2025

Google Faces Antitrust Trial Over Browser Dominance

Google faces the possibility of regulatory action, with a judge set to hold a trial next week on the DOJ's request to make Google sell its Chrome browser and take other measures to end its dominance in online search. The ruling is seen as a "major inflection point" for Google and the tech sector, highlighting US courts' willingness to entertain "aggressive structural remedies" in antitrust cases. This could increase regulatory risk premiums across major tech stocks, including Amazon and Meta, which operate similarly integrated ecosystems.

https://channelnewsasia.com/business/google-holds-illegal-monopolies-ad-tech-us-judge-finds-5072431

April 17, 2025

TSMC Sees Bright Future Amid AI Demand Boom

Taiwan Semiconductor Manufacturing Company (TSMC) is optimistic about its outlook despite tariff uncertainty, citing robust demand for artificial intelligence (AI) as a key driver. The company's shares have seen significant gains in recent times. TSMC is a leading manufacturer of semiconductors and has been expanding its presence globally.

https://www.livemint.com/companies/company-results/tsmc-upbeat-on-outlook-as-robust-ai-demand-offsets-tariff-uncertainty-11744879890512.html

April 16, 2025

AMD Predicted to Surpass Nvidia in Growth

AMD's Q1 2025 data center revenues are expected to dip around 7% quarter-over-quarter, but one investor, The Asian Investor, predicts a surprise to the upside. They believe AMD is set for a massive up-surge in 2025 due to its new AI accelerators, including the Instinct MI350 and MI400, which could turbocharge revenues and help the company pull even with or surpass Nvidia's data center growth. Despite tariffs being a risk, The Asian Investor thinks strong execution on AMD's growth strategy will give it resilience, and rates AMD stock a Strong Buy, citing a 52% upside for the year ahead.

https://markets.businessinsider.com/news/stocks/looks-like-a-steal-says-investor-about-amd-stock-1034587494

April 16, 2025

Apple Enhances Siri with AI Training System Upgrade

Apple is introducing its latest AI training system, integrated into iOS and iPadOS 18.5, as well as macOS 15.5, in an effort to improve Siri's performance amid technical issues. The company has been promoting differential privacy methods since iOS 10's launch in 2016, which it claims can protect user data by injecting random information into a dataset, preventing individual connections.

https://en.tempo.co/read/1997291/apple-claims-it-can-train-ai-without-compromising-users-personal-data