Mon Sep 25

News For Today

  • What Is Bhashini, India’s AI-Powered Language Tool

  • Meta Launches Its Answer To ChatGPT

  • AI Will Change Dating

  • A Startup is using AI to Personalize Clothing

What Is Bhashini, India’s AI-Powered Language Tool: The Indian government is advocating for indigenous tech alternatives like Bhashini, an AI-powered translation platform supporting 10 Indian languages. This move aligns with the National Language Translation Mission and the broader Digital India initiative, which has a robust budget of Rs 14,903 crore. Bhashini aims to democratize information by overcoming linguistic barriers in India, a nation with 19,500 languages and dialects but only 10% fluent English speakers. Bhashini complements global models like OpenAI's ChatGPT by focusing on Indian languages. PM Narendra Modi highlighted Bhashini at the G20 Digital Economy Working Group Ministers Meet, emphasizing India's tech prowess. The vision extends beyond Bhashini, aiming for a broader digitalized India through initiatives like modernizing the National Knowledge Network and promoting cybersecurity education. Bhashini's future involves a unified architecture, inviting public contributions to refine the platform.

Meta Launches Its Answer To ChatGPT: Meta has launched its own AI assistant for its platforms, WhatsApp, Instagram, and Messenger, positioning itself in the AI chatbot market alongside OpenAI's ChatGPT and Anthropic’s Claude. This general-purpose assistant can assist in group chats, answer queries, and with its partnership with Microsoft’s Bing, provide real-time web results. Unique to Meta's AI is its "/imagine" feature, which generates images similar to OpenAI’s DALL-E. Ahmad Al-Dahle, Meta’s VP of generative AI, mentioned the assistant is based on principles from Llama 2, their latest quasi-open source model. To make it more conversational, they refined its data sets and increased its context window for better user interaction. Meta is also introducing 28 AI characters, including celebrity-themed ones. These characters animate subtly during conversations, offering a more immersive experience. Meta has taken steps to prevent the AI from generating harmful content, spending significant time testing and refining its responses.

AI Will Change Dating: Bumble CEO Whitney Wolfe Herd envisions AI enhancing online dating. Speaking at the Code Conference, Wolfe Herd highlighted that Bumble has been using AI-driven matching algorithms for years, promoting compatibility and safety. While the company isn't interested in creating virtual partners, Wolfe Herd sees AI as a tool to alleviate user pain points and increase compatibility. She envisions an AI matchmaker where users detail their relationship preferences, and the AI converses with other users' matchmakers to assess compatibility. Only highly compatible matches would be introduced to users. Other potential AI applications include image recognition for shared interests and date planning. Wolfe Herd believes that AI can provide a premium, curated dating experience, hinting at a potential upscale offering in Bumble's future.

A Startup is using AI to Personalize Clothing: Blend, a UK startup, uses AI to offer tailored shopping recommendations for users. Unlike many retailers that use historic purchase data, Blend's AI-driven approach focuses on current user preferences, style evolution, and trends. Having participated in TechCrunch Disrupt 2023, Blend recently introduced its MVP app for a select user base and is now seeking seed funding. The app targets 18 to 34-year-old fashion-forward users and has partnered with over 250 retailers, including luxury brand Net-a-Porter. Blend aims to simplify online shopping by presenting users with the most relevant choices, personalized to their style, budget, and size. Its AI technology, based on transformer models, refines suggestions using user interaction data. As users engage more with the app, their recommendations become increasingly tailored. The app also aims to help users find the right fit by considering diverse body shapes and sizes, potentially reducing costly returns for retailers.