Meta is positioning artificial intelligence as both its primary growth engine and its frontline defence against platform fraud, as advertisers in Asia Pacific are now claimed to be achieving more than triple return on ad spend – even as scam enforcement scales into the hundreds of millions.
Meta Vice President of Asia Pacific Benjamin Joe said AI now underpins both performance outcomes and the integrity of Meta’s advertising environment.
“AI is at the core of everything we do for a global community of 3.4 billion people that are using our apps every day,” Joe said.
“For every $1 that an advertiser in APAC puts into meta, they get over three times that return.”
Those performance claims arrive as regulators worldwide sharpen expectations around scam prevention, data protection and AI transparency – pressures that are now reshaping the economics of digital advertising.
Meta Vice President of Asia Pacific Benjamin Joe
Scams move from safety issue to structural economic risk
Meta used this week’s Global Anti-Scam Summit in Washington, DC, to outline the scale of its enforcement efforts – and to make a broader point: that scam activity is no longer just a platform safety issue but a structural commercial threat.
For performance marketing, where credibility, speed and predictable audience behaviour underpin acquisition and lifetime value, sabotage of trust now directly suppresses conversion efficiency.
So far in 2025, Meta says it has:
• Removed more than 134 million scam ads
• Supported international investigations resulting in arrests
• Driven a more than 50% decline in user reports about scam ads across the past 15 months
Joe framed scams as an existential risk to the advertising system itself.
“Scams undermine the very trust that our business is built on,” he said.
“Scams don’t just harm individual victims; they undermine trust in our entire advertiser ecosystems. And as you know, this is the very foundation of Meta’s business model.”
For advertisers, that trust is now directly tied to media efficiency: hesitation weakens acquisition; weaker acquisition pressures loyalty economics, pricing power and growth models downstream.
From clean-up to prediction: AI becomes the new fraud firewall
Meta detailed a new suite of AI-driven detection systems designed to identify scam behaviour earlier and reduce downstream exposure.
These models are trained to flag:
• Emerging scam tactics
• Coordinated criminal behavioural patterns
• Network-level account anomalies across Facebook, Instagram and WhatsApp
One of the most commercially sensitive upgrades centres on facial recognition technology used to detect ads that misuse celebrity images or manipulated likenesses.
Meta says early testing more than doubled the volume of violative ads detected and removed.
This marks a strategic pivot from reactive content clean-up toward predictive, pre-exposure intervention, particularly as generative AI increases the speed, realism and scalability of fraud.
Targeting the criminal infrastructure behind scams
Beyond individual ads, Meta is now prioritising enforcement at the network level.
In the first half of 2025, the company says it:
• Detected and disrupted nearly 12 million accounts tied to what it describes as “criminal scam centres,” particularly in Southeast Asia
• Supported the US Department of Justice’s Scam Centre Strike Force and the FBI in dismantling a network linked to the Tai Chang compound in Myanmar, resulting in the takedown of 2,000 Facebook accounts
• Shared intelligence with Singaporean authorities, leading to the dismantling of an illegal online gambling network tied to digital fraud
These actions sit alongside more than 60 lawsuits Meta has filed in recent years targeting impersonation networks, account-takeover schemes, and coordinated abuse.
From social discovery to AI-guided decision-making
By 2026, Meta expects product discovery to increasingly begin on social – a creator video, a friend’s post, a Reel or a live stream – before AI agents step in as decision-making intermediaries.
That journey now typically follows a new commercial loop:
Social inspiration – AI-powered judgement – purchase – AI-driven post-purchase creation – renewed discovery.
Meta is already testing virtual try-ons, allowing users to see products on themselves before purchasing – designed to lift confidence at the moment of transaction.
Meanwhile, Meta Advantage+ automates end-to-end campaign optimisation, compressing the execution layer of performance marketing into real-time AI.
Messaging becomes the new storefront
Commerce is also consolidating inside chat environments. “Messaging is now the primary storefront on WhatsApp, Messenger and Instagram,” Joe said.
“APAC businesses are using AI agents to handle sales, support and loyalty at scale.”
Rather than pushing consumers between apps and websites, Meta’s strategy is to enable discovery, enquiry, and purchase within a single continuous conversation.
A real-world example comes from Philippines-based medical apparel brand White Coat Manila, which deployed business AI on Messenger to answer customer questions and guide users directly to cart – delivering a 20% drop in support costs and a 5% uplift in customer conversations.
The AI-driven creator loop
Creators are increasingly operating as product discovery infrastructure rather than just distribution channels.
The Facebook Affiliate Partnership with Shopee, now live across Southeast Asia, Taiwan and Brazil, allows creators to convert content directly into sales using in-platform affiliate links – collapsing content, commerce and conversion into a single performance loop.
Video, meanwhile, is now the default commercial language across mobile-first APAC markets.
Meta is piloting tools that let creators embed product links directly into Instagram Reels, while nearly two million advertisers are now using generative AI video tools to create custom ads, deploy AI-generated music, and optimise for multiple audiences and languages.
Live shopping, AR try-ons, and short-form video are now driving direct revenue, not just awareness, pushing video firmly into the performance media stack.
White Coat Manila
Cross-border commerce becomes the APAC growth engine
Asia Pacific is increasingly operating as the core engine room of global cross-border e-commerce.
As geopolitical and economic pressures reshape traditional trade routes, intra-APAC trade is accelerating, with AI-driven automation reducing friction across creative, budget allocation, placement and localisation.
Singapore-born furniture brand Castlery used Meta’s automation tools to expand into Australia, while Korean beauty label ANUA leveraged Reels and creator-led Partnership ads to crack the US market.
In both cases, automation reduced the complexity typically associated with cross-border scaling.
Agencies, automation and the changing value chain
While buying is increasingly compressed into automated systems, Joe argued agency value is not shrinking – it is shifting.
“We actually believe that the large advertisers who use creative agencies will continue to do that with the advantage that AI tooling will actually turbocharge those creative development processes.”
Execution is becoming machine-led. Strategy, creative systems, behavioural insight and brand architecture remain stubbornly human.
Meta is now prosecuting a single strategy across six converging fronts simultaneously:
• AI as a revenue engine
• A fraud suppression system
• A regulatory compliance buffer
• A media automation layer
• A creator-commerce infrastructure
• And a cross-border trade accelerator
Its assertion that advertisers are achieving 3x ROI while removing more than 134 million scam ads in a single year underscores the new reality of performance marketing: Trust is no longer moral capital – it is now economic infrastructure.