Meta Deploys Enhanced AI Scam Protection and Financial Ad Verification in Poland

Meta has launched an enhanced anti-impersonation AI system in Poland

Meta announced on August 28, 2026 that it is deploying an enhanced AI system in Poland to detect and limit scams that impersonate celebrities and other public figures. The company says Poland is among the first group of European countries receiving this newer protection.

This is a released regional safety-product rollout, not a new foundation model. Meta does not disclose the model architecture or enough evaluation detail in the announcement to independently benchmark the system.

Meta says its existing advanced AI systems already analyze scam and impersonation activity at scale. According to the company, the technology has helped reduce user reports involving the most frequently impersonated public figures by more than 80%. That figure is Meta-reported and should not be interpreted as an independently audited reduction in all scam activity.

Financial advertisers targeting Poland will face universal identity checks

Alongside the AI rollout, Meta is expanding advertiser verification for financial services. The company says 100% of financial-services advertisers directing ads to people in Poland will be required to verify their identity.

The timing matters: Meta says the rollout of this universal financial-advertiser verification requirement will be completed over the coming weeks. The requirement is therefore announced and in deployment, but should not be described as fully completed on August 28.

Meta positions the verification change as a second layer of defense around a category frequently abused by fraud operators. AI detection can identify suspicious content and impersonation patterns, while advertiser identity checks make it harder for bad actors to operate behind unverified financial-ad accounts.

Meta reports fewer scam-ad reports in Poland

Meta also published Poland-specific enforcement figures. It says the rate at which people reported scam advertisements in Poland fell 83% from July 2024 through June 2026. From July 2025 through June 2026, Meta says it removed 137,000 scam ads in Poland, with more than 88% removed before anyone reported them.

These are platform-reported operational metrics rather than independent measurements of the total amount of fraud experienced by Polish users. A lower report rate can reflect better prevention, enforcement and user experience, but it does not by itself establish the prevalence of scams across the wider internet.

Why this rollout is useful for AI safety and platform-security teams

The update shows how large consumer platforms are increasingly combining AI classifiers with identity and policy controls rather than relying on content detection alone. Impersonation scams are adversarial: attackers continuously change wording, imagery, accounts and destinations to evade static rules. A layered system can use AI to identify suspicious patterns while verification raises the cost of repeatedly creating deceptive financial-ad campaigns.

For trust-and-safety, ad-tech and fraud teams, the important development is not the headline percentage reduction alone. It is Meta's decision to move an enhanced impersonation-detection system into an early European market while simultaneously tightening advertiser identity requirements in a high-risk category.

What is released versus still upcoming

Released now: Meta says the enhanced anti-impersonation AI system is being launched in Poland on August 28, 2026.

Still rolling out: mandatory identity verification for all financial-services advertisers targeting Poland is expected to finish over the next several weeks.

The announcement does not establish that scams have been eliminated, does not provide a public model card or independent evaluation of the enhanced AI system, and does not justify extrapolating Meta's Poland-specific enforcement metrics to every country or platform.