Meta’s Digital Hall of Mirrors: When the Oversight Board Celebrates Manipulated Reality
In a decision that highlights the evolving challenges of content moderation in the age of artificial intelligence, Meta’s oversight board recently upheld the company’s choice to keep manipulated videos on Facebook despite their deceptive nature. The incident involved a post from the Philippines in which the video had been altered to falsely depict a politician in a dangerous situation promoting gambling. As reported by Engadget, the committee argued that while the content was misleading, it did not violate Meta’s policy on manipulative media, which focuses narrowly on AI-mediated alterations that make people say things they didn’t say. The ruling highlights the board’s nuanced approach in prioritizing freedom of expression over outright takedowns, despite what critics have denounced as a loophole for disinformation.
The Oversight Board, often likened to a Supreme Court of Meta platforms, was established in 2020 to review controversial content decisions on Facebook, Instagram, and Threads. The Board of Directors, made up of experts from a variety of fields including law, journalism and human rights, operates independently but receives funding from Meta. In this case, the manipulated video, which was edited using non-AI techniques, showed celebrities appearing in fabricated scenarios. The committee noted that Meta should have labeled it “high risk” to warn users, but ultimately supported keeping it online and emphasized the importance of political debate in the context of a democracy.
This is not a special case. Recent deliberations by the board have repeatedly revealed deficiencies in Meta’s handling of altered content. For example, in a June 2025 ruling reported by The Hindu, the board overturned Meta’s decision to leave AI-manipulated videos alone, calling Meta’s policy “incoherent.” This pattern suggests a tension between Meta’s desire for streamlined moderation and the Board’s desire for stronger safeguards against misinformation, especially during elections and crises.
Navigating the policy maze
Meta’s Manipulated Media Policy was introduced in 2020 and refined amid growing concerns about deepfakes, and primarily targets content that uses AI to fabricate audio. But as AI tools become more accessible, the lines between traditional editing and advanced operations are blurring. The Oversight Board’s decision in the Philippine case criticized Meta for failing to apply the “high risk” label, which could have mitigated potential harm without censorship, as detailed in a public statement. The recommendation echoes previous calls for transparency, including a 2024 assessment by the board that classified Meta’s rules as insufficient for an election year.
Industry insiders say Meta’s approach stands in sharp contrast to competitors such as X (formerly Twitter) and TikTok, which have introduced extensive labeling requirements for tampered content. Posts on X, including posts from users discussing Meta’s policies, reflect public frustration with sentiments ranging from condemning lax enforcement to defending free speech. The Guardian reported that one thread highlighted how Meta’s reluctance to remove content during the 2024 UK riots led to criticism from its board of directors, who accused Meta of making hasty changes to policy with little regard for the impact on society.
The board’s influence extends beyond individual cases. Since its founding, it has issued more than 200 decisions and numerous policy recommendations, many of which have been adopted by Meta, such as increasing transparency in content removal. The Oversight Board’s Wikipedia entry describes meetings with whistleblowers like Francis Haugen, who in 2021 exposed internal documents revealing that Meta prioritized engagement over safety. This context lends depth to the recent ruling and suggests that the board is attempting to balance Meta’s commercial interests with its ethical obligations.
Ripple effects on global elections
As the world grapples with the proliferation of AI-generated content, the impact of the Board’s decision extends far beyond the Philippines. Manipulated media has been used as a weapon to sway public opinion in the lead-up to major elections, including those in 2024 and 2025. The Associated Press reported in February 2024 that the board urged Meta to reconsider its “disjointed” policy on deepfakes and warned of the risk of disinformation. But the latest endorsement of manipulated video posts shows reluctance to expand removal criteria, potentially leaving the platform vulnerable to coordinated campaigns.
Critics, including digital rights groups, say this stance emboldens bad actors. An April 2025 Reuters article detailed the board’s rebuke of Meta’s policy overhaul, which eased restrictions on topics like immigration and gender identity and prioritized user retention over fact-checking. Regarding X, these decisions are often discussed as evidence of Meta’s bias against controversial content that boosts engagement metrics, with users citing the example of the Indian elections where proxy pages allegedly violated policies to influence the outcome.
Conversely, tech industry advocates see the board’s approach as a safeguard against overreach. The Board aims to foster open dialogue by focusing on intent and circumstances (such as whether an operation incites violence). This philosophy is consistent with Meta’s human rights policy, which is based on international standards such as the Universal Declaration of Human Rights and is referenced in the Board’s opinion. However, as a Springer analysis published in the journal Minds and Machines in 2022 points out, while the expansion has closed some gaps, the limited scope of the board, which initially only reviewed takedowns and did not decide to keep content, is at issue.
Challenges in enforcement and transparency
Enforcing these policies at scale remains Meta’s Achilles heel. Automated systems with billions of users handle most moderation, but subtle changes often fall flat. The Council’s recent call on Meta to address “information asymmetries” in conflict zones, as covered by Forbes magazine in November 2025, underscores this. During the Syrian crisis, uneven content removal created disparities and favored certain stories. In the case of manipulated videos, the commission recommended improved labeling to inform users, and Meta has partially implemented this, but critics say it falls short.
Public sentiment toward platforms like X amplifies these concerns. Posts from surveillance advocates praise the board’s proactiveness, such as dismissing meta about AI content, while others criticize perceived inconsistencies. For example, the decision to leave alone a post accusing Philippine politicians of corruption that used images of family members was upheld despite privacy concerns, as stated on the Oversight Commission’s website. This reflects a broader discussion. Should platforms pay attention, or should they be the ones expressing it at the wrong time?
Mehta’s reaction to the board’s recommendations has been mixed. The company’s Transparency Center lists the changes it has implemented, including strengthening fact-checking partnerships, while a 2025 Engadget article criticizes uneven AI moderation as “disjointed and unfair.” Insiders whisper that internal pressures, from advertiser demands to regulatory oversight, are complicating reforms. For example, the European Union’s Digital Services Act mandates tighter controls on disinformation, and Meta is under global pressure to adjust its policies.
Evolving standards in the post-truth era
In the future, the role of oversight committees may expand as advances in AI further blur reality. Recent news from Tech Edition in June 2025 criticized Meta for its handling of AI-manipulated content, echoing the board’s own frustrations. By insisting on the label while supporting the Philippine video, the commission has shown it prioritizes education over erasure, a strategy that could have implications for other platforms.
However, challenges still exist. Meta’s funding of the board raises independence issues, but its bylaws ensure operational autonomy. The 2025 Religion Unplugged analysis called on Meta to reduce the information gap in disputes based on board decisions. Regarding X, conservative voices have accused Mehta of liberal bias, citing human trafficking policies that allegedly prioritize profit over protection.
Ultimately, these rulings force consideration of digital ethics. As manipulated content proliferates, the board’s flawed but positive decisions could shape how society navigates truth in the age of synthetic media. Industry experts expect further policy adjustments, driven by board pressure and external regulation, to close loopholes without stifling speech.
The path forward for platform accountability
Meta and its interaction with the oversight board exemplifies the tech industry’s self-regulatory experiment. As in the June 2025 gambling promotion case reported by MarketScreener, the board has reversed decisions, but its influence is advisory. Meta’s report says the recommendations have an adoption rate of about 80%, but critics are calling for a binding mandate.
The global context adds a layer. In India, X-threads violate Mehta’s policies by claiming tolerance for ruling party ads. Whistleblower testimony continues to come under increased scrutiny in the United States. The commission’s mission, as stated on the site, is to improve the treatment of meth for communities around the world, a goal tested by cases like the Philippines video.
As AI evolves, we expect the debate to intensify. Boards seeking consistency could lead to comprehensive reforms to prevent platforms like Facebook from becoming echo chambers of deception. For now, the decision to leave manipulated content online serves as a wake-up call to the tightrope of moderation.
