AI Deepfake Rumors About Metro Collapse, a Woman Detained for Gaining Traffic

With the rapid development of artificial intelligence technology, some lawbreakers have begun to use AI tools to fabricate false information in order to gain traffic. This behavior seriously undermines the online ecosystem and has a negative impact on the normal social order.

Creating Panic by Deceiving the Public

Recently, at Jiangpu Road Station on Line 18 of the Shanghai Metro, a small amount of wall paint fell off from the ceiling of the transfer passage. There were no casualties at the scene. The metro staff immediately set up warning signs, initially judged that the falling paint was caused by water leakage, and completed preliminary repairs.

However, Sun, from another province, saw this news and, in an attempt to attract traffic, used an AI tool to generate a false video claiming "a large area of the ceiling collapsed inside the Shanghai metro station." She then posted the video on her WeChat video account, causing panic among netizens and a negative social impact.

Caught by the Police

With close cooperation from public security organs in other provinces, the rail transit police quickly identified the person who committed the violation, Sun, and arrested her. Currently, Sun has been administratively detained by the police for suspected creation and spread of online rumors.

At present, using AI to generate false texts, images, audio, and video has become a new trend in the spread of online rumors. Such information is highly realistic and misleading. The general public should be cautious when facing such online information, and do not believe or spread rumors.

From Painting Cats to Understanding the Body: Midjourney Crosses into Healthcare, Launches Full-Body Ultrasound Scanner

Midjourney enters the healthcare field by collaborating with Butterfly Network to launch its first hardware, Midjourney Scanner. This full-body ultrasound scanner uses a ring-shaped sensor array to extend AI image generation capabilities to quantitative perception of human body structures, marking its transition from creative imagery to a new field of medical diagnosis.

Anthropic Urgently Halts Claude Agent SDK Billing Change, Heavy Users Temporarily Maintain Existing Subscription Limits

Anthropic suddenly halted the planned June 15 billing adjustment for the Claude Agent SDK. The original plan was to unify external SDK usage under API rates, distinguishing it from the web interface and official CLI. Subscribers would only receive monthly equivalent credits. The decision has been postponed, maintaining the current pricing, allowing developers and heavy users to continue using the previous model.

AntTech Launches Agentar Financial Industry Intelligent Expert Group Covering Ten Core Financial Positions

On June 16, AntTech launched the Agentar Financial Intelligent Expert Group at the main forum of the Financial Exhibition. Targeting institutions such as banks, securities companies, and insurance companies, it covers core scenarios such as wealth management, risk control, and marketing, driving the transformation of financial AI from tool-based applications to position-level intelligent agents. The expert group includes ten financial digital experts and 300+ industry intelligent agents, with each expert corresponding to a complete job role, no longer limited to single-task execution.

WordPress VIP Releases AI Survival Report: Over 80% of Consumers Do Not Fully Trust AI-Generated Content, 42% Lower Trust Due to Lack of Source

According to the WordPress VIP report, consumer trust in AI continues to decline: 60% are annoyed by brands using artificial intelligence for marketing, 86% do not fully trust AI, and 73% believe the internet lacks more warmth. 42% of the audience have lower trust in content without clear sources, while it is easier for brands to access AI citations, but winning trust is more difficult.

WeChat Pay Launches AI Toolkit 2.0: Not Only Understands 9 Languages, but Also Reduces Token Consumption by Half!

WeChat Pay launches the AI Access Toolkit 2.0 version, which is upgraded based on the original product and user feedback, focusing on security, efficiency, and native smoothness. The new version supports 9 languages and conducts native semantic adaptation for WeChat Pay business concepts, avoiding machine translation errors, helping global developers complete integration in their native language without barriers.