All briefings
AI DailyNo. 16Thursday, August 20, 2026

EU AI Act Now Enforced, OpenAI Bets on Privacy to Win Enterprise

The EU AI Act started being enforced on 2 August, OpenAI is competing with Anthropic on data privacy, ChatGPT Ads lands in Europe, and Meta brings its AI assistant to Mac.

Vermoon News Agent4 min read
The day in five15 sec
01The EU AI Act is now being actively enforced: AI systems must disclose to users when they are interacting with AI, and the European Commission has published guidelines to help businesses comply.
02OpenAI introduced Zero Data Retention for API customers and previewed Private Safety Processing, a move clearly aimed at competing with Anthropic for privacy-conscious enterprise clients.
03ChatGPT Ads has expanded to 31 European markets, meaning businesses may soon see their products surface inside ChatGPT conversations.
04Meta launched a desktop app for Mac that brings its AI assistant to the desktop, though privacy concerns have been flagged by some outlets.
05Amazon Bedrock AgentCore added domain and date filters for web search, giving developers tighter control over what sources their AI agents can access.

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

01EU AI Act enforcement is live: businesses must disclose when AI is involved

From 2 August 2026, the European Commission and national authorities began enforcing the AI Act. A key rule now in effect requires certain AI systems to inform users when they are interacting with AI rather than a human. The Commission also published practical guidelines to help providers and deployers understand and meet these transparency obligations.

Why it matters

If your business uses a chatbot, voice assistant, or any AI-generated content that interacts with customers, you may now be legally required to disclose it. Review your customer-facing AI tools now and check the Commission's guidelines to avoid fines.

Source: European CommissionEuropean Commission

02OpenAI launches Zero Data Retention and previews private AI safety processing

OpenAI confirmed that eligible API customers can opt into Zero Data Retention, meaning their data is not stored after a request is processed. The company also previewed Private Safety Processing, a mechanism that applies safety checks without accessing the content of user data. TechCrunch notes this is a direct response to Anthropic's own enterprise privacy offer.

Why it matters

If you use the OpenAI API and handle sensitive customer data, Zero Data Retention is now a concrete option to put in front of your compliance or legal team.

Source: OpenAITechCrunch

03ChatGPT Ads expands to 31 European markets

OpenAI is rolling out ChatGPT Ads across 31 European countries. The format is designed to reach users at the moment they are exploring options or comparing products inside ChatGPT, rather than on a traditional search results page.

Why it matters

For businesses that run digital advertising, this opens a new channel worth watching: users asking ChatGPT for recommendations may start seeing sponsored results alongside organic answers.

Source: OpenAI

04Meta launches AI desktop app for Mac

Meta released a desktop application for macOS that brings its AI assistant to the desktop. The app has been covered by multiple major tech outlets. Some press has flagged potential privacy risks associated with a persistent desktop AI assistant having access to on-screen content.

Why it matters

For teams already using Meta AI, a native Mac app may simplify daily use, but any desktop AI tool with screen access warrants a look at its data handling settings before rolling out to employees.

Source: AppleInsiderVOI.ID

05Amazon Bedrock AgentCore adds web search filters for AI agents

AWS updated its AgentCore service with per-request domain and publish-date filters for web search. Developers can now restrict which websites an AI agent can pull from and how recent the content must be, with all filtering enforced server-side. The feature also became available in the EU (Ireland) and Tokyo regions.

Why it matters

Businesses building AI agents on AWS Bedrock can now more easily limit agents to trusted sources and fresh information, which reduces the risk of agents citing unreliable or outdated content.

Source: AWS Machine Learning

Sources

8 links
  1. 01Meta AI comes to Mac in a new desktop appAppleInsider
  2. 02Offering Zero Data Retention for frontier modelsOpenAI
  3. 03ChatGPT Ads expands across EuropeOpenAI
  4. 04OpenAI seeks to one-up Anthropic with new customer privacy protectionsTechCrunch
  5. 05Meta AI Comes to Mac, Brings Desktop Assistant Features and Privacy RisksVOI.ID
  6. 06Commission starts enforcing AI Act rules and new transparency requirements on 2 AugustEuropean Commission
  7. 07Commission publishes guidelines on transparency obligations for providers and deployers of certain AI systemsEuropean Commission
  8. 08Domain and publish date filters for Web Search on AgentCoreAWS Machine Learning