google, amd and openai Updates

The artificial intelligence landscape is rapidly evolving, demonstrating significant advancements across various sectors, from enhanced privacy measures to sophisticated video generation and critical infrastructure applications. Google recently introduced Private AI Compute, a system designed to ensure user data privacy when utilizing its Gemini cloud AI models. This technology leverages specialized AMD-based hardware to encrypt and isolate memory, making user information temporary and employing third-party relays to mask IP addresses. An independent audit by NCC Group confirmed the system's adherence to privacy and security protocols. AI's progress in video generation is particularly striking, with models like MiniMax, Google's Veo 3, and OpenAI's Sora now producing remarkably lifelike results, a stark contrast to the "horrifying" outputs of just two and a half years ago. This rapid improvement is already seeing commercial application, with companies like Coca-Cola using advanced AI for advertisements. Concurrently, the cybersecurity sector is bolstering its defenses with AI, as evidenced by a busy November 2025 for acquisitions. Palo Alto Networks, for instance, plans to acquire Chronosphere for $3.35 billion to enhance its AI resilience, joining other major players in integrating AI into security solutions. Regulatory bodies are also beginning to address AI's impact, with New York becoming the first state to pass a law against "personalized pricing." This legislation mandates that retailers inform customers when an algorithm determines their price, aiming to prevent unfair practices based on personal data. Ethical concerns surrounding AI deployment continue to emerge, as seen with Flock Safety, a US surveillance company, which uses gig workers in the Philippines to train its AI models by reviewing and classifying footage from American streets. This practice, intended to improve license plate recognition for law enforcement, raises significant privacy worries, especially since police often use Flock cameras without warrants. Beyond these developments, AI is finding crucial applications in public safety and economic growth. California utility companies are deploying AI-powered solutions from firms like Overstory and Rhizome to prevent costly wildfires by identifying and managing high-risk trees near power lines. On the hardware front, Japanese company Murata is significantly increasing its production of electronic components like capacitors and RF filters, anticipating a 30% annual growth in AI accelerator production until 2030. Internationally, Vietnam is making ambitious strides, aiming to become a top three AI research and development hub in Southeast Asia by 2030, focusing on creating its own AI products, including a large language model tailored to Vietnamese culture. Even in politics, an organization called The Independent Center is using AI to identify congressional districts where independent candidates could potentially disrupt the traditional two-party system.

Key Takeaways

• Google launched Private AI Compute, utilizing AMD-based hardware, to ensure user data privacy for its Gemini cloud AI models.
• Flock Safety employs overseas workers, many in the Philippines, to train its AI surveillance systems, raising privacy concerns regarding footage from US streets.
• New York passed the first state law against AI-driven "personalized pricing," requiring retailers to disclose algorithmic price determination.
• AI-generated video technology, including models from OpenAI (Sora) and Google (Veo 3), has advanced significantly, producing lifelike results.
• Palo Alto Networks plans to acquire Chronosphere for $3.35 billion to enhance its AI resilience and cybersecurity capabilities.
• Japanese company Murata is increasing production of electronic components, anticipating a 30% annual growth in AI accelerator production until 2030.
• California utility companies are deploying AI tools from firms like Overstory and Rhizome to prevent wildfires by managing vegetation near power lines.
• Vietnam aims to become a top three AI research and development leader in Southeast Asia by 2030, focusing on indigenous AI products.
• Cybersecurity firms are actively acquiring other companies to boost their AI security tools and automate defenses.
• An organization is using AI to identify congressional districts suitable for independent political candidates, aiming to reshape US politics.

Flock uses overseas workers for AI surveillance

Flock, a company that makes automatic license plate readers and AI cameras, uses workers from overseas, many in the Philippines, to train its AI. These workers review and classify footage from US streets, identifying vehicles and even listening for sounds like gunshots. This practice raises privacy concerns because police use Flock cameras in many US communities, often without warrants. An online panel showing worker details was removed after a media inquiry.

Flock Safety uses Philippine workers for AI surveillance

Flock Safety, a US surveillance company, hires gig workers in the Philippines to train its AI models. These workers annotate vehicle footage from American streets to improve license plate recognition for law enforcement. This practice, revealed by leaks, brings up privacy worries, even though Flock says it anonymizes sensitive data. The company aims to stop crime with its cameras, which are in thousands of communities. Critics are concerned that AI trained by distant workers could flag unusual movements as suspicious.

Vietnam aims to be a top AI leader in Asia

Vietnam plans to become a major AI hub in the Asia-Pacific region, as stated by Minister Nguyen Manh Hung at Vietnam International Digital Week 2025. The country has a national strategy to be among the top three in Southeast Asia for AI research and development by 2030. Vietnam focuses on creating its own AI products, including a large language model that understands Vietnamese culture. UNESCO has praised Vietnam's strong commitment and rapid progress in AI research, with scientific output growing significantly.

Google launches Private AI Compute for secure data

Google introduced Private AI Compute, a system designed to use its Gemini cloud AI models while keeping user data private. This technology uses special AMD-based hardware to encrypt and isolate memory, ensuring user information stays secure. The system also makes data temporary, meaning it only keeps information long enough to answer a user's request. Google also uses third-party relays to hide users' IP addresses, adding another layer of privacy. An independent auditor, NCC Group, confirmed the system's design meets privacy and security rules.

New York tackles AI personalized pricing first

New York became the first state to pass a law against "personalized pricing," where businesses use AI and personal data to set prices. This new rule requires retailers to inform customers when an algorithm determines their price. The law aims to prevent stores from unfairly raising prices based on a customer's past spending habits. While some groups find the law too broad or too narrow, experts agree it is a significant step in regulating AI. This move could start a trend for other states to address similar AI pricing practices.

Murata boosts production for AI hardware growth

Japanese company Murata is increasing its production of key electronic parts like capacitors and RF filters, betting on the long-term growth of AI hardware. The company expects strong demand from AI servers and advanced cars, predicting AI accelerator production to grow 30% yearly until 2030. Murata also sees its RF filter demand rising by 17% annually through 2028 and car electronics increasing significantly. To support this, Murata plans to expand its production capacity and develop new technologies. The company also aims for carbon neutrality by 2050, showing a commitment to sustainable growth alongside the AI boom.

Will Smith spaghetti videos show AI video progress

AI-generated video technology has advanced greatly in just two and a half years, as shown by videos of Will Smith eating spaghetti. In 2023, early AI models created unrealistic and "horrifying" images, making the "spaghetti test" a benchmark for improvement. Since then, models like MiniMax, Google's Veo 3, and OpenAI's Sora have produced much more lifelike results. This rapid progress shows how quickly AI video generation is improving. Despite ongoing legal challenges from Hollywood over content use, companies like Coca-Cola are already using advanced AI for their advertisements.

Cybersecurity firms boost AI security with acquisitions

November 2025 was a busy month for cybersecurity companies buying other firms to strengthen their AI security tools. Major players are adding AI, monitoring, and risk management features to their products. For example, Palo Alto Networks plans to buy Chronosphere for $3.35 billion to improve its AI resilience. Other notable deals include LevelBlue acquiring Cybereason, Safe Security buying Balbix, and Zscaler purchasing AI security startup SPLX. These acquisitions show a clear industry trend towards using AI to automate and improve cybersecurity defenses.

California utilities use AI to fight wildfires

California utility companies are turning to artificial intelligence to help prevent costly wildfires. Since burying power lines is too expensive, companies like Overstory and Rhizome offer AI-powered solutions. Overstory uses AI to find and manage trees that pose a fire risk near power lines, helping utilities like Edison International avoid huge damage claims. Rhizome's gridFIRM tool helps utilities decide the best ways to reduce fire risks, considering local weather and populations. National Grid already uses this technology in New York and Massachusetts. These AI tools are crucial as climate change makes wildfire prevention more challenging.

Organization uses AI to reshape US politics

The Independent Center, an organization, is using artificial intelligence to change American politics. It aims to identify congressional districts where independent candidates could win. This strategy seeks to disrupt the current two-party system in the United States.

Sources

• Flock Uses Overseas Gig Workers to Build Its Surveillance AI
• Flock Safety Outsources AI Surveillance Training to Philippine Gig Workers
• Vietnam set to become a regional AI powerhouse
• Private AI Compute Enables Google Inference with Hardware Isolation and Ephemeral Data Design
• NY Becomes First State to Target AI-Driven Personalized Pricing
• Murata Bets Big On AI Hardware’s Long-Term Upgrade Cycle
• Then vs. now: AI videos of Will Smith eating spaghetti show just how advanced the tech has gotten
• Cyber M&A Roundup: Cyber Giants Strengthen AI Security Offerings
• California utilities try AI to combat costly wildfire liabilities
• This organization is trying to use AI to reshape American politics