Yazılar

Google Launches Multiple Open-Source Translation Models Following ChatGPT Translate

/in /tarafından

Google launches open AI translation models challenging ChatGPT Translate

Google has continued its aggressive push in the artificial intelligence (AI) space in 2026. Following a series of initiatives including a partnership with Apple, the launch of new shopping tools, the introduction of Personal Intelligence in Gemini, and the integration of a chatbot into its Trends website, the company is now focusing on the open community. Its latest move comes with the release of TranslateGemma, a set of multilingual AI models designed for translation across a wide range of languages, supporting both text and image (input only) modalities.

In a recent blog post, Google announced three different variants of the TranslateGemma models. These AI models are available for download through Google’s Hugging Face listing and Kaggle, and can also be accessed via Vertex AI, the company’s cloud-based AI hub. Google has released the models under a permissive licence, enabling both academic and commercial use cases, which encourages broader adoption and experimentation by developers and enterprises alike.

The TranslateGemma models come in three sizes: 4B, 12B, and 27B parameters. The smallest, 4B model, is optimized for mobile and edge deployment, while the 12B variant targets consumer laptops, balancing performance and efficiency. The largest, 27B model, provides maximum translation fidelity and can run locally on a single Nvidia H100 GPU or TPU, making it suitable for high-demand environments and research applications.

With this release, Google is aiming to democratize AI-powered translation and provide the developer community with robust, versatile tools. By making TranslateGemma models open-source and easy to deploy across different devices, Google is reinforcing its commitment to accessible AI and strengthening its position in the multilingual AI ecosystem. The move also highlights the growing importance of translation technology in enabling global communication and cross-cultural collaboration.

EU’s AI Code of Practice for Firms Likely Delayed Until End of 2025

/in /tarafından

The European Commission announced on Thursday that the Code of Practice designed to help companies comply with the EU’s Artificial Intelligence Act (AI Act) may only come into effect by late 2025. This code aims to guide thousands of businesses on meeting the new AI regulations, especially for general-purpose AI (GPAI) models like OpenAI’s ChatGPT, Google’s, and Mistral’s AI systems.

Background and Delay Calls

  • The Code of Practice was originally slated for publication on May 2, 2025, but its release has been delayed.

  • Major tech companies, including Alphabet (Google), Meta, and European firms such as Mistral and ASML, alongside some EU governments, have requested postponements due to the lack of clear compliance guidelines.

  • The European AI Board is currently debating the timeline, with end of 2025 under consideration for full implementation.

Voluntary but Important

  • Signing up for the Code is voluntary, but companies that refuse will not gain the legal certainty given to signatories.

  • The Code will clarify the expected quality standards AI service users can demand, reducing risks of misleading claims by providers, according to Nick Moës, Executive Director of AI advocacy group The Future Society.

  • The Code also involves oversight by legally mandated authorities to assess AI service quality.

EU’s Position and Industry Reaction

  • Despite calls for delay, the Commission insists it remains committed to the AI Act’s goals of harmonized, risk-based AI regulations and market safety.

  • Critics, such as campaign group Corporate Europe Observatory, accuse Big Tech of using delay tactics to weaken crucial AI safeguards.

Enforcement Timeline

  • The AI Act’s rules on GPAI models become legally binding on August 2, 2025, but enforcement will begin only a year later, on August 2, 2026, for new models entering the market.

  • Existing AI models have until August 2, 2027, to comply fully with the regulations.

Google Names DeepMind CTO Koray Kavukcuoglu as Chief AI Architect to Lead AI-Powered Product Development

/in /tarafından

Alphabet’s Google has appointed Koray Kavukcuoglu, the chief technology officer of its DeepMind AI lab, as its new chief AI architect and senior vice president, according to an internal memo from CEO Sundar Pichai. Kavukcuoglu will relocate from London to California and report directly to Pichai, while continuing his role as DeepMind CTO under CEO Demis Hassabis.

In this expanded leadership position, Kavukcuoglu will drive faster integration and iteration of Google’s cutting-edge AI models into its wide array of products, aiming to increase efficiency and seamless adoption as generative AI gains mainstream traction.

The move comes as Alphabet faces mounting pressure to justify its projected $75 billion AI investment this year by translating breakthroughs into tangible financial returns. Google must balance these efforts with maintaining profitability amid competition from rival AI developers and heightened antitrust scrutiny.

Google recently unveiled an AI subscription service priced at $249.99 per month targeting power users, alongside demonstrations of new AI-enhanced products like smart glasses during its May I/O conference. CEO Pichai emphasized that the ongoing generative AI expansion complements rather than replaces traditional online search.

Additionally, Google has formed a notable partnership with OpenAI — one of its biggest AI competitors — by agreeing to supply cloud computing resources to OpenAI’s operations, highlighting the evolving dynamics in the AI sector where collaboration and competition coexist.

This strategic leadership appointment signals Google’s intent to accelerate the transition into a new phase of AI platform development and adoption.