Meta expands business AI agents across WhatsApp, Instagram, and Messenger
Meta is rolling out AI agents for businesses across WhatsApp, Instagram, and Messenger, pushing customer support and sales tools deeper into its apps.
AI News from July 5, 2026
Latest Artificial Intelligence Updates and Stories
July 5, 2026
Doubao and Qwen reportedly set July 15 shutdown for personalized AI agents under Chinese rules
Doubao and Qwen will reportedly close personalized AI agent features on July 15 as Chinese regulation tightens oversight of user-created AI services.
Mystery model ‘Hunter Alpha’ is being linked to DeepSeek V4, but public evidence remains thin
A mystery AI model called Hunter Alpha is being linked to DeepSeek V4, highlighting how opaque model testing and anonymous releases shape AI competition.
Alibaba and Tencent appear to back Kling AI, signaling cooperation over direct competition in China’s video model race
Reports say Alibaba and Tencent backed Kling AI, a move that suggests China’s AI leaders may prefer platform bets over direct video model rivalry.
ByteDance and Alibaba reportedly curb AI agent features in China as regulators tighten focus on autonomous systems
ByteDance and Alibaba are reportedly disabling some AI agent functions in China, signaling stricter limits on autonomous AI products.
Mistral CEO Arthur Mensch warns enterprises that closed AI models expose business process data
Mistral CEO Arthur Mensch urged companies to avoid dependence on closed AI models, arguing data control and model ownership matter in enterprise AI.
KAIST says AI agents carry a hidden energy burden, highlighting a new cost center beyond model training
KAIST reports a newly identified hidden energy cost in AI agents, raising fresh questions for enterprise AI deployment and agent design.
DiscoBench finds AI search agents break down on ambiguous queries because they don’t ask users to clarify
A new benchmark from Tencent Hunyuan and Tsinghua University suggests AI search agents fail on ambiguity, not search itself, limiting real-world reliability.
Z.ai’s GLM-5.2 appears aimed at Anthropic-style coding and agent workflows, but the evidence is still thin
Z.ai appears to be positioning GLM-5.2 against Anthropic-class coding and agent models, highlighting rising open-weight pressure in enterprise AI.
Study links advanced AI features to sharply higher power use as model compute races ahead
New reporting suggests advanced AI systems can draw far more electricity than basic chatbots, raising new cost and infrastructure questions for AI buyers.
Mistral AI reportedly seeks $3.5 billion as it signals a new open-weights model for this summer
Mistral AI is reportedly seeking $3.5 billion and says an open weights model is due this summer, a move that could reshape open model competition.
Ads