New AI Models, Video Tools, and Growing Privacy Pressure

Google, Microsoft, and OpenAI have all launched new AI models and tools in recent days, while Perplexity faces a serious privacy lawsuit. This development highlights both rapid innovation and increasing pressure around trust and data use in the AI market.

Google Announces Gemma 4 Open AI Models

Google has released Gemma 4, an update to its open AI models, now available under the Apache 2.0 license. The update includes four models optimized for local use, giving developers greater flexibility to run advanced AI systems on their own machines.

The models are designed for lower latency and better performance with local processing, which is becoming increasingly important as more users seek to reduce reliance on the cloud. This trend is also seen in other models, such as in our review of the Granite 4.0 model.

Source: Ars Technica

NVIDIA Optimizes Gemma 4 for Local AI

NVIDIA has partnered with Google to tailor the Gemma 4 models for its GPUs, enabling efficient operation of advanced AI models both in data centers and on powerful local machines. The optimization focuses on high performance and low response times.

This points to a clear trend of moving AI closer to the user, especially for applications requiring fast response or high data security. At the same time, this can reduce costs associated with cloud-based operations.

Source: NVIDIA Blog

Google Vids Gets AI Upgrade with Veo and Lyria

Google has updated its video platform Vids with new AI models Veo 3.1 and Lyria 3. The update enables generation of high-quality video and audio, including the use of controllable avatars.

This makes advanced video production more accessible to both businesses and individuals, lowering the barrier to producing professional content. The development builds on a broader push into AI video, as we also described in our review of how Google is heavily investing in AI video.

Source: Google Blog

Microsoft Launches Three New AI Models

Microsoft has introduced three new AI models for speech, image, and text generation. The models were developed by the company’s internal AI team and are part of a broader strategy to strengthen its own AI solutions.

The launch shows that competition in the AI market is intensifying, especially in multimodal systems. Several tech giants are now focusing on their own models to reduce dependence on external providers.

Source: TechCrunch

Perplexity Sued for Possible Privacy Violations

AI search company Perplexity has been sued for allegedly sharing user data with Google and Meta without consent. The lawsuit includes both subscribers and non-subscribers and raises questions about how AI companies handle sensitive data.

The case fits into a broader trend where trust in AI is under pressure, as we have also seen in our analysis of how AI usage is increasing but trust is declining.

Source: Ars Technica

OpenAI Acquires TBPN

OpenAI has acquired the technology podcast TBPN, a move that could strengthen the company’s communication and visibility in the public sphere. The acquisition gives OpenAI a direct channel to the tech community.

The investment may help shape how AI is discussed publicly, while giving the company greater control over its own narrative amid increasing competition and critical attention.

Source: TechCrunch

What Does This Mean?

AIny brief assessment: The AI market is clearly moving in two directions: more local and cost-effective AI, and increasing pressure on privacy and trust. At the same time, launches from Google and Microsoft show that competition is intensifying, especially in multimodal models. Cases like the Perplexity lawsuit could be decisive for how AI companies handle data going forward.

Read also: Perplexity Accused of Privacy Violations with Incognito Mode