Gambling Ads on Meta Hit Young Men 2.3x More

Cambridge Study Exposes Gambling Ad Bias on Meta Platforms

A new study from Cambridge University reveals that gambling advertisements on Meta's platforms — including Facebook and Instagram — reach young men at 2.3 times the rate of women, even when advertisers do not explicitly target male audiences. The research, which analyzed 411 ads from 88 licensed gambling operators in Ireland, raises urgent questions about how algorithmic ad delivery systems amplify harm among the most vulnerable demographics.

The findings arrive at a critical juncture in the global debate over AI-driven advertising, platform accountability, and the growing intersection of social media algorithms with public health outcomes. As regulators in the EU, UK, and US grapple with how to govern algorithmic systems, this study provides some of the most granular evidence yet that ad delivery algorithms can produce discriminatory outcomes — even in the absence of deliberate targeting.

Key Findings at a Glance

• Young men aged 25-34 were the most heavily exposed demographic, accounting for over one-third of all ad-reached accounts
• 12.6 million male accounts were reached across all 411 ads, compared to just 5.4 million female accounts
• 91 ads (22%) were explicitly targeted at men only — zero ads targeted women exclusively
• A single Betfair ad reached over 1.32 million unique accounts, equivalent to 26% of Ireland's entire population
• Ads targeting the 25-44 age group accounted for 59.4% of all reached accounts
• Among Irish men aged 25-34, 1.3% exhibit gambling addiction, compared to just 0.2% of women in the same bracket

Algorithms Amplify Risk Without Explicit Instructions

The most striking aspect of the Cambridge study is not that some gambling ads deliberately target men — that finding, while notable, is perhaps unsurprising. The far more consequential discovery is that Meta's algorithmic ad delivery system independently skews gambling ad exposure toward young male users, regardless of advertiser intent.

This phenomenon, sometimes called 'algorithmic discrimination' or 'proxy targeting,' occurs when machine learning systems optimize for engagement and conversion metrics. Because young men are statistically more likely to interact with gambling content, the algorithm learns to serve these ads disproportionately to that demographic. The result is a feedback loop: the algorithm identifies high-engagement users, serves them more ads, and reinforces the very behavior patterns that make this group most susceptible to harm.

Unlike traditional advertising — where a billboard or TV spot reaches a broad, undifferentiated audience — programmatic ad delivery on social media platforms uses hundreds of behavioral signals to determine who sees what. This makes the platform an active participant in targeting decisions, not merely a passive channel. The Cambridge researchers argue this distinction is critical for regulators to understand.

Ireland as a Case Study With Global Implications

The research focused on Ireland for several practical reasons: the country has a well-documented gambling landscape, a relatively small population that makes comprehensive analysis feasible, and a regulatory framework that requires gambling operators to hold licenses. But the implications extend far beyond Irish borders.

Ireland's gambling addiction statistics paint a concerning picture. Among men aged 25-34 — the exact demographic most heavily reached by these ads — 1.3% exhibit problem gambling behavior. For women of the same age, the figure drops to 0.2%, a 6.5x disparity that the ad delivery patterns appear to mirror and potentially exacerbate.

The scale of individual ad campaigns is also remarkable. A single advertisement by Betfair, one of the world's largest online betting exchanges, reached more than 1.32 million unique accounts in Ireland. In a country of approximately 5.1 million people, that means a single gambling ad touched roughly 1 in 4 Irish residents with a Meta account. This level of penetration would be virtually impossible through traditional media channels and underscores the unprecedented reach that social media platforms afford to gambling operators.

The Broader AI and AdTech Accountability Debate

This study feeds directly into one of the most heated debates in the technology industry: who is responsible when algorithms cause harm? Meta has historically maintained that its ad delivery system is a neutral tool, and that responsibility for ad content lies with advertisers. However, a growing body of research — including this Cambridge study — challenges that framing.

Several key questions emerge from the findings:

• Should platforms be held liable for algorithmic amplification of harmful content, even when advertisers do not explicitly request discriminatory targeting?
• Do current ad transparency tools (such as Meta's Ad Library) provide sufficient visibility into how ads are actually delivered versus how they are targeted?
• Should gambling ads be subject to the same algorithmic restrictions that some jurisdictions have proposed for political advertising?
• How should regulators account for the difference between advertiser intent and algorithmic outcome when crafting policy?

These questions are not merely academic. The EU's Digital Services Act (DSA), which took full effect in 2024, requires very large online platforms to assess and mitigate systemic risks, including risks to public health. The Cambridge findings suggest that gambling ad delivery on Meta's platforms may constitute exactly the kind of systemic risk the DSA was designed to address.

Compared to previous studies that examined gambling advertising in broadcast media or static online placements, this research breaks new ground by quantifying the amplification effect of algorithmic delivery. Earlier work by researchers at institutions like the University of Bristol and Carnegie Mellon had documented similar patterns in housing and employment ads on Facebook, leading to a landmark $5 million settlement with the US Department of Housing and Urban Development in 2022. The gambling domain, however, has received far less regulatory scrutiny.

What This Means for Platforms, Regulators, and Users

For Meta and other social media companies, the study represents a growing body of evidence that their ad delivery systems produce outcomes that may conflict with both regulatory expectations and their own stated commitments to responsible advertising. Meta's advertising policies prohibit targeting gambling ads at minors, but the research suggests that age-gating alone is insufficient when algorithmic delivery disproportionately reaches the most at-risk adult populations.

For regulators, the findings highlight a gap between existing advertising rules — which typically focus on content and explicit targeting parameters — and the reality of algorithmic ad delivery. Traditional regulatory frameworks were designed for a world where advertisers chose their audience; in the algorithmic era, the platform's AI makes many of those decisions autonomously.

For users, particularly young men in the 25-34 age bracket, the research confirms what many may have intuitively suspected: their social media feeds are saturated with gambling content to a degree that far exceeds what other demographics experience. This has direct implications for mental health and financial well-being.

Practical steps that stakeholders might consider include:

• Platforms implementing algorithmic audits specifically for gambling ad delivery patterns
• Regulators mandating demographic impact assessments for gambling ad campaigns on social media
• Gambling operators voluntarily adopting balanced-reach requirements to avoid disproportionate exposure
• Researchers conducting similar analyses in larger markets such as the UK, US, and Australia
• Users leveraging ad preference tools to reduce gambling ad exposure in their feeds

Looking Ahead: Regulation, Litigation, and Technical Solutions

The Cambridge study is likely to fuel several parallel developments in the coming months. In the regulatory arena, Irish and EU authorities may cite the findings as evidence supporting stricter controls on algorithmic advertising in gambling. The Irish government has been developing comprehensive gambling regulation, and this research provides empirical ammunition for those advocating tighter rules.

In the legal sphere, the documented disparity in ad delivery could form the basis of discrimination or consumer protection claims, particularly if plaintiffs can demonstrate that algorithmic amplification contributed to gambling harm.

On the technical front, the study underscores the need for more sophisticated ad delivery auditing tools. While Meta's Ad Library provides some transparency, it does not fully reveal the algorithmic decisions that determine actual ad distribution across demographics. Researchers have long called for API access and more granular data to enable independent oversight.

The intersection of AI-powered advertising and public health is an area that demands immediate attention. As ad delivery algorithms grow more sophisticated and platforms expand into new markets, the patterns documented in this study are unlikely to remain confined to Ireland's gambling landscape. They represent a systemic challenge that will require coordinated responses from technologists, policymakers, and civil society alike.