Meta faces $25 billion fine, backtracks on scraping terms, and restricts adult creators as moderation failures persist

Record $25 billion fine for data misrepresentations

A New Mexico judge ordered Meta to pay $25 billion for misleading the public about its handling of Facebook user data, marking the largest fine ever imposed on the company. The lawsuit stemmed from Meta’s role in the Cambridge Analytica scandal, where data from millions of users was harvested for political consulting without consent. A jury found Meta violated New Mexico’s consumer protection laws over 40 million times, with the judge describing the violations as "willful" and spanning more than 11 years.

The original penalty was calculated at $219 billion but was reduced to $25 billion. Meta plans to appeal the decision, arguing the fine is excessive and detrimental to its business operations in New Mexico. The company has also faced multiple settlements in 2026, including an $18 billion deal with California over allegations of designing addictive platforms and failing to protect children from psychological harm.

Backlash forces reversal on scraping terms

Meta updated its commercial terms of service in October to include language permitting the use of content from business accounts for AI training and other purposes. Legal experts warned the provision could allow Meta to scrape publisher websites with minimal restrictions, as robots.txt guidelines are not legally binding and often ignored by AI companies.

After backlash from publishers and digital rights groups, Meta removed the word "train" from the updated terms and clarified that the changes apply only to its ads and business products. Andy Stone, a Meta spokesperson, insisted the update did not change how the company uses web crawlers. However, critics argue the revised terms still grant Meta broad permissions to scrape content, raising concerns about consent and compensation for creators and publishers.

AI sexting allowed while human creators face restrictions

OnlyFans creators and adult content professionals are criticizing Meta for permitting its AI agent, Muse, to engage in explicit sexting while simultaneously restricting human creators on Instagram and Facebook. Muse, available to users aged 18 and older, generates explicit text but not images or videos. Creators argue this double standard unfairly disadvantages human performers, who face account suspensions or restrictions for similar content.

Vanniall, an OnlyFans creator with over 100,000 Instagram followers, described the policy as hypocritical, stating: "It's not fair that they get to pull money out of people with their sexually active AI agent, but they won't let real people, real adults, engage in any sort of sensuality or sexuality." Adult performers, including queer and trans creators, also report frequent account removals or restrictions on Meta’s platforms, while competitors like X and Bluesky allow sexually explicit content.

The discrepancy highlights broader concerns about AI displacing human creators, particularly those in marginalized communities, while being trained on their content without consent or compensation.

Content moderation failures persist

Meta’s oversight board identified "recurrent enforcement failures" in content moderation during the US midterm elections and Brazil’s 2022 presidential election. The failures included over-enforcement that suppressed political speech and under-enforcement that undermined electoral integrity. In Brazil, Meta was ordered to prove it treated presidential candidates equally after allegations surfaced that the company suppressed content from Luiz Inácio Lula da Silva’s campaign.

The board urged Meta to improve transparency and consistency in moderation practices, particularly during elections. Separately, Meta temporarily deactivated an Instagram account, "No Kings," after automated systems targeted high-frequency posters, apologizing for the error afterward.

Privacy vulnerabilities in AI platform

Meta suspended a feature in its Muse AI platform after researchers discovered a privacy vulnerability that exposed over 6,800 shared conversations. The flaw allowed access to sensitive information, including meeting times with AI labs, loan applications, and internal company documents. Meta paused the sharing functionality while reviewing the issue, highlighting ongoing risks in AI deployment.

Meta’s $26 billion privacy penalty bill

Since 2020, social media platforms have accumulated over $26 billion in privacy fines and settlements globally, with Meta’s Facebook and Instagram accounting for 91% of the total. The US imposed $20.9 billion in fines, primarily for data-processing violations and exposing minors to addictive features. European regulators levied $5.16 billion, with Ireland issuing 94% of the continent’s fines due to Meta’s regional headquarters being based in Dublin.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

Meta’s recent legal and policy challenges underscore the growing risks for tech platforms operating at scale. The $25 billion fine—even if reduced on appeal—signals that regulators are increasingly willing to hold companies accountable for historical misconduct, particularly when it involves data privacy and public misrepresentation. For founders and operators, this case highlights the long-term costs of inconsistent transparency and the need to align internal practices with public statements.

The backlash over Meta’s scraping terms reveals a broader tension between AI development and content ownership. While Meta framed the changes as limited to business products, publishers and creators viewed them as an overreach, forcing the company to retreat. This episode serves as a warning for startups building AI tools: vague or expansive terms of service can trigger resistance, especially when they appear to prioritize data acquisition over consent and fair compensation.

The disparity in Meta’s treatment of AI-generated versus human-created adult content exposes inconsistencies in moderation policies, particularly for marginalized creators. Platforms like X and Bluesky are gaining traction by permitting explicit content, while Meta’s restrictions risk alienating creators who feel censored. Startups in the creator economy may find opportunities to differentiate by adopting clearer, more equitable policies—especially as AI tools blur the lines between human and synthetic content.

Meta’s moderation failures and AI privacy lapses also demonstrate the pitfalls of over-reliance on automation. The company’s struggles to balance enforcement during elections, combined with vulnerabilities in its AI platform, underscore the need for robust safeguards and human oversight. For founders building AI products, this reinforces the importance of testing for privacy risks and designing systems that can adapt to nuanced moderation challenges.

Related on Lazyfounder

• Nvidia unveils OpenShell for AI agent safety and DSX POD for faster data centre deployment
• Arm’s AGI CPU Outperforms AMD in Agentic AI Benchmarks, Signal65 Reports
• Apple and Meta partner to bring Muse Spark Open AI model to Mac Studio
• Canada’s encryption bill sparks tech exodus threats