Meta & Children's Online SafetySenate Judiciary Committee: Written Statement of Jason Sattizahn
Raw text
WHISTLEBLOWER AID Senate Judiciary Committee Subcommittee on Privacy, Technology and the Law Written Statement of Jason Sattizahn September 9, 2025 Chairman Blackburn, Ranking Member Klobuchar, and Members of the Subcommittee. Thank you for having me here. I am here to discuss Meta's manipulation of research to cover up dangers facing billions across Meta's products, and of particular concern, the millions of children using Meta's virtual reality products. I also want to recognize and thank the five other past and current Meta employees – all researchers like me who worked directly on creating Meta's products – who made the brave decision to be a part of the disclosure that brought us here. My name is Jason Sattizahn. Growing up in the nineties in the middle of Missouri, I saw both the value and problems that fast advances in technology brought the world. After earning a PhD in Integrative Neuroscience, I realized I wanted to use my research experience to make these technologies and products better for people that use them. Most recently, I spent six years as a researcher at Meta. I worked in some of Meta's most sensitive spaces, tasked with understanding users, their needs, and using this to try and make their products safer. I am here today because it is evident that Meta consistently chooses profit over safety for its users. I am not the first to discuss this, as repeated whistleblowers have shared Meta's reckless disregard for users. However, in the wake of past whistleblowers, Meta has chosen to ignore the problems they created and bury evidence of users' negative experiences, rather than build safer experiences. I worked at Meta from 2018 to 2024. During these six years, I witnessed data scandals, multiple disclosures about Meta's disregard for user safety and children's mental health, and mounting public pressure against Meta. I saw the company respond to these pressures by deliberately compromising internal processes, policies, and research to protect company profits over users. WHISTLEBLOWER AID Written Statement of Jason Sattizahn (Continued) During my first role at Meta, I led integrity research for Facebook Marketplace. The data was clear - Marketplace causes suffering for users, including financial loss from stolen or counterfeit items, and personal safety issues ranging from being sexually propositioned by strangers to physical assaults and attempted kidnapping. My time on Marketplace was my first exposure to how Facebook de-prioritized safety to boost user engagement. Simple safety investments such as not allowing people to message strangers with a single click were flatly rejected because product teams were afraid to do anything that could possibly decrease engagement - the metric largely determining success and bonuses. It was around this time I first saw Facebook make false statements to Congress. Contrary to Mark Zuckerberg's previous testimony before Congress, my research enables the company to estimate the overall rates of various types of harm occurring to Facebook Marketplace users. For instance, we can pair just one of the many studies I lead on this topic, estimating the extent of various harm based on estimates of Facebook's ~1 billion active Marketplace shoppers. For example: • • • • • • 9% of Marketplace users reported at least one negative experience in the past month (est. 90 million Facebook users per month). 0.5% of Marketplace users reported scams/counterfeit goods per month (est. 5 million reports of scams/counterfeit goods per month). 0.44% of Marketplace users reported harassment per month (est. 4.4 million reports of harassment per month). 0.3% of Marketplace users reported seeing drugs/alcohol per month (est. 3 million reports of seeing drugs/alcohol per month). 0.2% of Marketplace users reported stolen goods per month (est. 2 million reports of stolen goods per month) 0.16% of Marketplace users reported physical threats/assault per month (est. 1.6 million reports of physical threats/assault per month) • 0.4% of Marketplace users reported seeing inappropriate images in Marketplace posts (est. 4 million reports of inappropriate photos per month) When Facebook launched “Faith” efforts to connect religious communities during the Covid lockdown era, I joined because I was scared that Facebook's pattern of disregard for safety would surface in this already fraught context. Sure enough, my research on the use of Facebook for Faith showed large numbers of harms to Facebook users, including: 2 WHISTLEBLOWER • • AID Written Statement of Jason Sattizahn (Continued) Facebook's own systems for detecting and removing inappropriate content online were faulty, often detecting religious content as “bad” and removing it; this caused multiple faith-groups to see Facebook as religiously persecutorial Facebook users' religious spaces would contain harassing, distracting, or otherwise distressing individuals in ways that harm faith-related experiences; this distressing behavior was at times intentional, and other times related to users not having common understanding of proper behavior in these spaces • Other Facebook Faith users experienced severe physical safety risks, including being lured into public and harassed/assaulted based on their faith Regardless, recommendations that I made such as creating well-understood community standards and expectations for behaviors in these spaces, and for the company to improve training we give operations teams to detect harm were shut down. Facebook instead focused on features to boost engagement, like an ability to respond to posts with a prayer hands emoji. The explicit reason given for my team focusing on new features rather than safety was because our team was expected to launch something to increase user engagement for the company rather than any other projects. This expectation was so strong that teammates I worked with on the faith team shared a deep fear that if they were not able to quickly build something to increase engagement, they could lose their jobs. In the Fall of 2021, Frances Haugen disclosed to Congress how Meta's products fuel mental health issues for teens, including body dysmorphia and self-harm. Meta's immediate response to Congressional concern was not to do the right thing, but rather, roll out new processes and policies to manipulate, control, and erase data . Researchers were directed how to write reports so that risk to Meta was limited, and internal workplace groups were locked down, making it nearly impossible to share data and coordinate between teams to keep users safe. Mark Zuckerberg disparaged whistleblowers, claiming past disclosures were “...used to construct a false narrative." Meta CTO Andrew Bosworth began arguing 1:1 with researchers that dared question the new restrictions placed on research. Despite Meta's attempts to prevent researchers from collecting necessary insights, the research we were able to do continued to show the dangers of Meta's products on users. This only highlights the sheer scale, severity, and prevalence of harm occurring to those using Meta's products. 3 WHISTLEBLOWER AID Written Statement of Jason Sattizahn (Continued) What Meta had in effect created was a funnel of manipulation, the ability to review, control, and have retroactive power over all stages of research at the company. This enabled them to prevent, manipulate, or erase any research that could potentially reflect negatively on Meta or its products. This control included: • • • Surveillance by the legal team. Researchers in sensitive spaces (i.e. integrity, safety, Youth) were given a legal partner they were required to consistently share research plans, data, and reports with for approval at all stages of work Restricting data collection. Ο Ο Ο Meta controlled topics researched, specific questions asked, and even types of research methods allowed, all to avoid negative findings Meta monitored research when performed, in order to enact stop work orders and halt research if it was seen as too risky or sensitive Meta demanded third-party contactors be used to house reports of user harm so Meta could claim ignorance to “knowing” these findings Whitewashing research findings. Meta would retroactively review reports to soften statements, edit evidence-backed claims by researchers, or erase findings entirely from reports • Research destruction. Meta would have evidence entirely destroyed (i.e. recordings, notes, etc.) so that the only evidence that really existed of sensitive problems was the knowledge living within researchers' themselves Research isolation. Meta would silo or segregate research they felt was too "risky” or sensitive, effectively limiting or disabling the ability for researchers to share their knowledge with others in the company • In early 2022, I moved to Meta's Reality Labs to lead integrity research and to help improve the product so people were safe while they used Meta's VR headsets. Generally, virtual reality allows someone to wear a headset and experience an alternate reality where they play games, watch movies, and socialize with others. For Meta, VR is designed to push socialization above all, as Meta saw this as a path to unbridled increases in user engagement and profit. The company invested billions, integrated social media like Instagram into headsets, and even rebranded as “Meta” to align with the future of the company. From my first days in Reality Labs, Meta leadership and legal teams were in complete control of the research I was conducting. The research was crucial since this was a largely untested technology, but I soon learned that Meta had no interest in VR safety unless it could drive interactions, and thus, profit. Despite their efforts to control 4 WHISTLEBLOWER AID Written Statement of Jason Sattizahn (Continued) research, my first studies showed strong evidence of the harm happening to users of Meta Virtual Reality: • Just shy of half of all VR users (44%) reported harm in the past three weeks when using their virtual reality headset • • These harms ranged from things like rudeness (39% of VR users) and entering others' personal space (19% of VR users) to racism (17%) and physical sexual advances (14%) Of people reporting harm in VR, an overwhelming 70% reported that it occurred in VR experiences meant for socializing – the same apps that Meta is actively and excessively promoting to people using their headsets In this same research, I discovered that certain harms were worse for women using Meta VR, particularly sexual harms such as harassment, molestation, or worse: • 26% of women reported experiencing sexism – compared to 14% of men when using virtual reality • • 23% of women reported experiencing verbal sexual advances (e.g. sexual solicitation) – compared to 9% of men – when using virtual reality 17% of women reported experiencing other sexually inappropriate behaviors (e.g. broadcasting audio of pornography, broadcasting audio of actual sexual acts) - compared to 7% of men – when using virtual reality For these women, this harm they faced in Meta VR also caused worse emotional and psychological harm, as well: • Harm in VR caused negative anxiety or stress for 47% for women, compared to only 23% of men • Harm in VR caused negative embarrassment for 39% for women, compared to only 17% of men When I discovered these disturbing findings of what VR users were facing, Meta demanded that I not write any reports that showed that the harm was specific to Meta or Meta Virtual Reality, despite the data showing that was true. Rather, all reports were to show that this was industry-wide harm and not specify Meta. It was also after these studies that Meta made changes to my research, completely removing questions about psychological or emotional harm to their users. Their reasoning? They didn't “want” the data, as they decided it was too “risky” for them to 5 WHISTLEBLOWER AID Written Statement of Jason Sattizahn (Continued) know. Effectively, Meta had hard evidence that VR was causing their users and women – psychological harm, but doctored research to prevent gathering that information in the future. After Meta's VR sales were banned in Germany for two years over concerns about how Meta treats user data, in 2022, Germany allowed sales to resume. When I was asked to perform research in Germany, I understood that Meta was trying to show that their VR headsets were safe for German users. However, when our research uncovered that underage children using Meta VR in Germany were subject to demands for sex acts, nude photos, and other acts that no child should be exposed to, Meta demanded I erase any evidence of such dangers. During my time working on Virtual Reality, I repeatedly saw leadership decide to invest resources on boosting social engagement at the direct expense of safety investments. At one point, Meta decided to allow children and people across devices, like phones or computers, to play together in their flagship VR app – Horizon Worlds. Meta had performed no research on how to make this experience safe for children or adults in this new space. As I pushed back and insisted on due diligence to prevent bad actors from hurting people, leadership explicitly told me that Meta had decided to launch without any safety research to “boost user engagement” in Horizon Worlds. Meta's corruption of research for their own protection isn't limited to a certain product like VR, but rather, it is what defines the past, present, and future of Meta's products. In my work at Meta, I collaborated across multiple areas and saw how Meta had the same purposeful avoidance in addressing user safety across all their future-facing technologies. For example, my discussions with Meta's wearable tech (glasses, wrist technologies) teams showed me they were avoiding investments to proactively address user safety. Even Meta's own law enforcement team admitted internally that they investigated cases such as the use of Meta's Ray-Ban glasses in the 2025 New Years terrorist attack in New Orleans, but that due to internal restrictions, the team was not allowed to share information with product teams that would help Meta build this future hardware safely. Despite Mark Zuckerberg's bold-faced lie to Congress that teams are not instructed to focus on engagement, that is exactly how the company operates. The pursuit of user engagement was the value that permeated Meta's every decision. For years, our work force was told that the directive to prioritize growth and engagement came from Mark Zuckerberg himself. 6 WHISTLEBLOWER AID Written Statement of Jason Sattizahn (Continued) Given how zealously Meta undermines reality, they will do their best to publicly undermine our disclosures with a counter narrative reflecting their unique version of "truth". They'll offer statistics without context or scale to prop up meager safety efforts. As they have done repeatedly, they may just lie to the government. They'll clutch their pearls, claiming the whistleblowers are “taking research out of context”. Ironically, in coming forward to protect Meta's users, I'm still doing the job that Meta originally hired me for. Unfortunately, my own experience and those of current Meta employees part of this disclosure show that Meta's focus on engagement over safety continues to this day, driving their vision for their future products, such as Al, wearables, and the focus on miniaturizing new technologies to integrate with people's everyday lives. Meta is incapable of change without being forced by Congress. Meta's leadership will not willingly change their mindset and insatiable appetite for profits and growth at all costs. We are here because when Meta has been given repeated and frankly unearned opportunities to improve their product and take responsibility for suffering they have created, they have chosen to double down, deflect and deceive. To hide evidence. To punish those doing the right thing. To disregard and disrespect government bodies like yours instead of working with you to make positive change. I believe that Congress can use what I and many others have disclosed to enact legislation and take action to compel Meta's leadership to create the safe and accessible technology we all deserve. Thank you again for the opportunity to be with you today and for your commitment to stop the deliberate harm for the millions of Americans who use this technology. 7