Welcome back to Daily Zaps, your regularly-scheduled dose of AI news ⚡️ 

Here’s what we got for ya today:

• 📧 Google adding search emails with AI to Gmail

• 🕴Hollywood wants to protect copyright from AI

• 🗣️ Human sounding text to speech (TTS)

• 🌐 Anthropic’s Claude chatbot adds web search

Let’s get right into it!

Sign Up | Advertise | Tools

BIG TECH

Google adding search emails with AI to Gmail

Google is enhancing Gmail’s search functionality with an optional AI-powered upgrade to help users find emails more efficiently. Instead of displaying results purely in chronological order, the improved search will also consider factors like recency, most-clicked emails, and frequent contacts. This update aims to surface the most relevant emails at the top of search results, saving time and improving search accuracy.

The new “most relevant” search results feature is now rolling out worldwide for users with personal Google accounts. It is available on both the web and the Gmail app for Android and iOS. However, Gmail’s traditional keyword-based search remains accessible, allowing users to toggle between chronological results and the AI-powered “most relevant” option.

ENTERTAINMENT

Hollywood wants to protect copyright from AI

More than 420 Hollywood actors, directors, and creatives, led by Natasha Lyonne, have signed an open letter urging the government to uphold copyright laws against AI companies like OpenAI and Google, which have proposed easing restrictions for AI training. The letter warns that such changes would harm the entertainment industry, which supports over 2.3 million jobs and contributes $229 billion in wages annually.

This push follows the 2024 SAG-AFTRA strike, where actors won protections against AI-generated replicas. California has since passed laws requiring consent for AI use in performances. Meanwhile, Vice President JD Vance has cautioned against AI overregulation, and President Trump has pledged significant investment in AI infrastructure while repealing prior AI safety measures, a move criticized by the ACLU.

STARTUPS

Human sounding text to speech (TTS)

Artificial intelligence startup Canopy Labs, aiming to create “digital humans indistinguishable from real humans,” has introduced Orpheus—a family of speech-generating large language models (LLMs) designed for high-quality text-to-speech (TTS) synthesis. Built on Meta’s Llama-3b architecture and trained on over 100,000 hours of English speech alongside billions of text tokens, Orpheus can generate natural, emotionally expressive speech, addressing the limitations of previous open-source TTS models.

It also features zero-shot voice cloning, an emergent capability from its extensive training corpus. Canopy Labs claims Orpheus enables real-time speech streaming with latencies as low as 25-50ms, thanks to a novel token sampling and decoding strategy. The company has released a three-billion-parameter model on GitHub and Hugging Face under the Apache 2.0 license, with smaller models (1B, 400M, and 150M parameters) to follow, making high-quality speech synthesis accessible even on resource-constrained devices.

STARTUPS

Anthropic’s Claude chatbot adds web search

Anthropic's AI-powered chatbot, Claude, now supports web search, a long-awaited feature currently available in preview for paid U.S. users, with broader access coming soon. Enabled via the Claude web app, this feature is limited to the latest model, Claude 3.7 Sonnet, and provides direct citations for fact-checking.

While it enhances Claude’s knowledge base with real-time insights, initial testing showed inconsistent activation for current events. This update brings Claude in line with competitors like ChatGPT and Google Gemini, despite Anthropic’s earlier stance on keeping it self-contained. However, the risk of misinformation remains, as studies show AI chatbots frequently provide incorrect or misleading responses.