Google Ads AI 2026: Disclosure Labels and MCP Server Reshape Ad Management

The world of AI-powered advertising is shifting rapidly. On July 13, 2026, two significant developments landed that every Google Ads advertiser needs to understand. Google is rolling out mandatory AI disclosure labels on ads, and Markifact launched a hosted Google Ads MCP server that lets AI tools like Claude and ChatGPT manage campaigns with built-in human approval. Here’s what these changes mean for marketers, regulators, and anyone running ads.

Markifact Launches Google Ads MCP Server for Claude and ChatGPT

The key change is that Markifact announced the global launch of its hosted Google Ads MCP server, enabling AI tools like Claude and ChatGPT to research, launch, report, and optimize campaigns with built-in human approval workflows. Unlike Google's own developer-focused MCP server, Markifact's version is designed for marketers and includes granular permission controls to prevent accidental changes. Markifact's official product page describes it as a "marketer-friendly alternative" with read-only query modes and approval gates for write operations.

The server connects via Google Ads API and OAuth, meaning it respects existing account permissions. Markifact hosts the connection layer, so advertisers don't need to set up their own infrastructure. The announcement on Business Insider notes that the server supports campaign creation, bid adjustments, budget changes, and performance reporting—all through natural language conversations with Claude or ChatGPT.

How the MCP Server Works with Human Approval

Markifact's server introduces a "human in the loop" safety layer. When an AI model requests a write operation—like changing a bid or pausing a campaign—the server pauses and sends a notification to the advertiser via email or dashboard. Only after explicit approval does the change execute. This addresses a major concern: AI making unauthorized changes to live ad spend. The server also logs every action for auditability.

For advertisers, this means they can ask Claude "Show me which campaigns have a ROAS below 3x" or ask ChatGPT "Increase the budget on our top-performing Shopping campaign by 15%." The AI pulls the data from Google Ads and, for read queries, returns answers immediately. For writes, it queues the change for human approval.

Comparison: Markifact MCP vs. Manual Google Ads Management

Feature	Markifact MCP Server	Traditional Manual Management
Campaign creation	Natural language prompt	Manual form entry
Bid adjustments	AI proposes, human approves	Manual change in UI
Performance reports	Instant conversational query	Build and export manually
Human oversight	Granular approval gates	Full control but slower
Infrastructure	Hosted, no setup needed	Must use Google Ads UI or API
Learning curve	Low (chat interface)	Moderate to high

This table shows that the MCP server doesn't remove human control—it shifts the work from manual clicks to conversational commands with a safety net.

Google Mandates AI Disclosure Labels on All Ads

Simultaneously, Google is rolling out mandatory AI disclosure labels across its ad ecosystem. Starting July 9, 2026, Google requires advertisers to label ads created or edited with generative AI. The labels appear in the "How this ad was made" section of My Ad Center on Search, YouTube, and Discover. Google's official announcement states that ads made with Google's own generative AI tools will be automatically labeled, while advertisers using third-party tools must manually enable disclosures.

According to TechCrunch, the move comes 24 days before the EU AI Act compliance deadline, suggesting regulatory pressure is the primary driver. The Silicon Report confirms that non-compliance could lead to ad disapproval or account suspension. The Verge adds that the labels will initially roll out globally, with visible overlays on ads in regions with stricter AI transparency laws like the EU, India, and New York.

How Disclosure Labels Work on Google Ads

Google's support documentation details the AI content label settings. Advertisers can specify whether an asset was AI-generated, AI-edited, or not AI-involved. The label appears in the My Ad Center panel, and in some regions, a "AI-generated" overlay appears directly on the ad creative. The system differentiates between "AI-generated" (substantially created by AI) and "AI-assisted" (human-created but edited with AI).

For example, if an advertiser uses Google's Performance Max with automatically generated assets, those will be automatically labeled. If an advertiser uses an external tool like Midjourney or an AI video editor, they must manually select the appropriate disclosure. The Search Engine Journal warns that mislabeling could be considered a policy violation.

Regional Variations and the EU AI Act

The EU AI Act, which comes into full effect soon, requires clear labeling of AI-generated content. Google's rollout timing—24 days before the deadline—is no coincidence. As PPC Land notes, the labels help advertisers comply without building custom solutions. In the EU, the overlay is mandatory. In other regions, the label is visible in My Ad Center but not always on the ad itself. Common Thread Collective explains that e-commerce brands using AI-generated product images or ad copy must now disclose, impacting how they test creatives.

What Advertisers Need to Do Now

Advertisers should take three steps:

1. Review all active ads for AI-generated or edited assets.
2. Enable AI content label settings in Google Ads for any third-party AI tools used.
3. Audit workflows to ensure human approval gates are in place for AI-generated creatives.

For those using the Markifact MCP server, the human approval workflow automatically aligns with Google's transparency requirements, as all changes are logged and can be associated with specific AI prompts. This makes audit trails easier.

The Broader Impact on AI Advertising

These two developments—Google's disclosure mandates and Markifact's MCP server—signal a maturing AI advertising ecosystem. On one hand, regulators demand transparency to prevent consumer deception. On the other, AI tools are becoming more powerful and accessible for campaign management. The tension between automation and accountability is being resolved through layered controls: AI proposes, humans approve, and disclosures inform.

Markifact's server is one of several new tools. Product Hunt shows other AI ad tools like Creatify's AI Media Buyer, but the MCP standard allows interoperability across AI assistants. As Google's Help documentation indicates, the company is investing in both AI capabilities for advertisers and transparency for users.

FAQ

Q: Do I need to label ads created by Google's own AI tools? A: No—Google automatically labels ads created with its generative AI tools (e.g., Performance Max automatically generated assets). You only need to manually disclose if you use third-party AI tools to create or edit your ad assets.

Q: What happens if I don't label an AI-generated ad? A: Google may disapprove the ad or suspend your account for policy violations. In regions like the EU, non-compliance could also lead to regulatory fines under the AI Act.

Q: Can Marcifact's MCP server help with AI disclosure? A: Indirectly—by logging all AI-driven changes, it provides an audit trail that can help you verify which assets were AI-generated, though you still need to apply the disclosure label manually in Google Ads.

Q: Will AI disclosure labels hurt ad performance? A: Initial data is limited, but Google's research suggests transparency can build trust. Some users may ignore labels; others may engage more with disclosed AI ads if the content is relevant.

Q: Is the Markifact MCP server free? A: The product page does not specify pricing publicly. It is a hosted commercial service likely involving a subscription or usage fee.