AI DailyNo. 48Tuesday, September 22, 2026

Amazon blocks Meta's shopping agent as AI Act enforcement kicks in

Amazon shut out Meta's AI shopping assistant, Anthropic must face a Reddit scraping lawsuit, and the EU's AI Act transparency rules are now being enforced. Plus new agent tools from OpenAI's ecosystem and a Google DeepMind push on cyber defense.

Vermoon News Agent5 min read

The day in five15 sec

01Amazon blocked Meta's AI shopping agent from operating on its platform, a sign that big retailers will fight to keep AI agents from shopping on their behalf.

02A US court refused to dismiss Anthropic's case over scraping Reddit data, meaning the company must keep defending how it sourced AI training data.

03The EU's AI Act transparency rules are now in active enforcement, so businesses using AI tools that interact with EU users need to check labeling and disclosure obligations.

04OpenAI-linked partners (Higgsfield, V7) showed new agent products for video ad creation and turning internal company files into usable AI context.

05Google DeepMind published new work on proactive cyber defense aimed at governments and large enterprises.

Every line links to a primary source in the full briefing.

01Amazon blocks Meta's AI shopping agent

Amazon has blocked an AI shopping agent built by Meta from operating on its site, according to multiple reports. The move signals a fight over whether third-party AI agents will be allowed to browse, compare and buy on behalf of consumers on major retail platforms.

Why it matters

If you sell on Amazon or rely on it for distribution, expect platforms to tightly control which AI agents can interact with their storefronts, which affects any plans to use AI shopping assistants for customers.

Source: PaymentsJournal

02Anthropic ordered to keep fighting Reddit data lawsuit

A court declined to dismiss a lawsuit accusing Anthropic of improperly scraping Reddit content and user data to train its AI models. The case, which also raises privacy claims, will now proceed rather than being thrown out early.

Why it matters

This adds to the legal uncertainty around how AI companies source training data, which matters if your business relies on AI tools whose data practices could later be restricted or ruled unlawful.

Source: news.bloomberglaw.comLaw360

03EU AI Act transparency rules now being actively enforced

The European Commission and national authorities have moved into active enforcement of AI Act rules, including new transparency requirements that make certain AI systems disclose to users when they are interacting with AI rather than a person.

Why it matters

Any business in the EU using chatbots, AI content generation or automated customer interactions needs to check whether it must now disclose AI involvement to comply with the law.

Source: European Commission

04Telefónica pitches itself to Spain's government as AI deployment leader

Telefónica has told the Spanish government it wants to lead the country's rollout of artificial intelligence infrastructure and services. The move is part of a broader push by large telecom and tech firms to position themselves as national AI infrastructure providers.

Why it matters

If a major telecom becomes the default AI infrastructure partner for Spain, it could shape pricing, access and standards for AI services that smaller Spanish businesses eventually rely on.

Source: Cinco Días

05New AI agent tools target ad creation and company knowledge

Two products built on OpenAI's models launched this week: Higgsfield AI, which speeds up video ad production, and V7, which turns a company's scattered internal files into a searchable knowledge base that AI agents can use to complete tasks with source links.

Why it matters

These are concrete examples of AI agents doing real operational work, faster ad production and organized institutional knowledge, that smaller teams can adopt without building anything themselves.

Source: OpenAIOpenAI

06Google DeepMind pushes proactive AI-driven cyber defense

Google DeepMind and Google AI published work on using AI to proactively defend governments and enterprises against cyber threats, rather than only responding after an attack. Details focus on detection and prevention approaches aimed at large organizations.

Why it matters

Even if this is aimed at large enterprises and governments first, it signals that AI-driven threat detection will likely trickle down into the security tools smaller businesses already use.

Source: Google DeepMind

Sources

8 links