AI companion safeguards tighten as New York targets emotionally manipulative chatbot features, the UK opens a broad parliamentary AI inquiry, and Deepgram acquires Daily to strengthen Pipecat’s open-source voice infrastructure. Meanwhile, Coveo and Digital Science expand agentic commerce and research tools, while lawmakers in New York and Australia raise the stakes for AI accountability. Scroll down to catch the signals that matter.

🧒 New York Draws a Hard Line on AI Companions

New York’s new law restricts unsafe AI companion chatbot features for minors, including harmful content and emotionally manipulative interactions. The measure requires operators to use permissible age assurance before providing covered features, creating a stricter design baseline than California’s broader governance framework.

AI companion companies now face child-safety obligations that begin before a conversation starts.

🛡️ Age Verification Moves Into the Chatbot Stack

New York’s companion-chatbot rules require operators to distinguish minors from adults before enabling restricted interactions. The approach connects age verification with safeguards against harmful content, emotional dependency, and other unsafe experiences, pushing platforms to treat identity assurance as a core product function rather than an optional compliance layer.

Age assurance is becoming foundational infrastructure for consumer AI.

🧠 Emotional Manipulation Becomes a Regulated Risk

New York’s restrictions target chatbot behaviors that could exploit children’s emotions or encourage unhealthy dependence. By limiting emotionally manipulative AI interactions, lawmakers are moving beyond content moderation toward regulating conversational dynamics, including how systems respond when young users seek comfort, validation, or constant companionship.

The next generation of AI safety rules will scrutinize behavior, not just output.

🏛️ UK Parliament Opens a Full AI Accountability Inquiry

A UK parliamentary committee launched an inquiry covering AI adoption, safety, governance, regulation, testing, and accountability. The broad remit signals that lawmakers are examining the entire deployment chain, from organizational incentives and technical evaluation to public protections and responsibility when automated systems cause harm.

The inquiry could shape how Britain balances AI adoption with enforceable oversight.

🛒 Coveo Buys Anglera for Agentic Product Data

Coveo acquired Anglera to add agentic product-data enrichment capabilities to its AI-powered commerce platform. The deal targets a persistent retail challenge: keeping product information accurate, structured, and useful for intelligent experiences. Better enrichment could help commerce agents interpret catalogs, answer questions, and support more relevant discovery.

Product data quality is becoming a competitive advantage for AI commerce.

✍️ Digital Science Adds Jenni AI to Research Infrastructure

Digital Science acquired Jenni AI, bringing its research-writing capabilities closer to published research data and infrastructure. The combination could help researchers move more smoothly between drafting, evidence discovery, and scholarly workflows while strengthening connections between generative writing tools and the underlying research ecosystem.

Research AI is shifting from standalone writing assistance toward connected scholarly infrastructure.

🎙️ Deepgram Acquires Daily to Advance Voice Agents

Deepgram announced its acquisition of Daily, the creator and maintainer of open-source Pipecat. The framework coordinates voice and video streams with speech recognition, language models, and speech synthesis, while reducing latency. Deepgram said Pipecat will remain open source, giving developers a broader foundation for expressive, real-time agents.

Open voice infrastructure could accelerate more capable and flexible AI agents.

⚡ UK Data Centres Get Strategic Grid Priority

The UK unveiled a Data Centre Covenant that prioritizes strategic AI facilities for electricity-grid connections while requiring community benefits. The policy links infrastructure access with public value, potentially accelerating critical compute development while pressuring operators to demonstrate local economic, environmental, and social returns.

AI infrastructure expansion will increasingly depend on both grid strategy and community consent.

🔍 RAISE Act 2.0 Targets Frontier AI Risks

New York lawmakers proposed RAISE Act 2.0, expanding safety, transparency, auditing, and enforcement requirements for major AI developers. The proposal includes reporting on automated AI research, rapid disclosure of serious risk changes, independent frontier examiners, and authority to restrict dangerous activity or deployment.

The bill would make frontier-model oversight a continuing state responsibility, not a one-time review.

🚫 Australia Moves Against Child-AI Emotional Intimacy

Australia’s House approved amendments targeting emotional intimacy between AI chatbots and children under digital duty-of-care legislation. The proposal treats simulated closeness as a safety issue in its own right, challenging platforms to prevent systems from encouraging dependency or presenting artificial relationships as substitutes for trusted human support.

Child-focused AI regulation is expanding from harmful content to synthetic relationships.

The global AI policy center of gravity is moving toward enforceable safeguards for children, infrastructure, and frontier systems.

Which of today's stories changes your roadmap? Tell us in one line.

Forward the signal: share your referral link

Reach data and AI leaders in this inbox: book a placement.