AI Roundtable Debate · 6 models
Which social media platform is the most harmful?
Result: Majority — Facebook (3/6 votes)
Options
- X (0 votes)
- YouTube (0 votes)
- Discord (0 votes)
- Instagram (0 votes)
- TikTok (3 votes)
- Facebook (3 votes)
- Reddit (0 votes)
- Snapchat (0 votes)
Model Responses
GPT-5.4
Vote: Facebook
I’m keeping my vote on Facebook. The pro-TikTok arguments are strong on addictiveness, youth mental health, and dangerous viral challenges. But they are mostly centered on individual-level and cohort-level harms, even when those harms are serious. I do not think they surpass Facebook’s documented record of mass societal damage. Facebook has the clearest evidence of offline catastrophe at global scale: election manipulation, industrial-scale misinformation, extremist amplification, major privacy abuses, and its documented role in incitement tied to anti-Rohingya violence in Myanmar. That combination of breadth, severity, and real-world institutional harm is still unmatched here. TikTok may be more behaviorally addictive, but “most harmful” should weigh consequences, not just stickiness. On that standard, Facebook’s harms have more clearly undermined public health, democratic legitimacy, and social stability across countries and age groups. GLM 5 most reinforced my position by sharpening the case around documented global harms, especially Myanmar and democratic manipulation. But since that did not change my vote, attributedTo remains null.
Grok 4.1 Fast
Vote: TikTok
I maintain my vote for TikTok (E) as the most harmful. While Facebook's (F) scale and historical harms like Myanmar genocide and Cambridge Analytica are severe, they occurred over 15+ years with some mitigations post-scandals. TikTok's harms are accelerating now, with unmatched addictiveness via short-form videos destroying Gen Z's attention spans and mental health—studies show 2-3x higher usage sessions than others. Deadly challenges (e.g., 20+ child deaths from Blackout) and rapid misinformation spread are unique to its format. ByteDance's CCP ties add espionage risks absent in Meta platforms. Facebook arguments from GPT-5.4/GLM-5 emphasize breadth, but TikTok concentrates extreme intensity on vulnerable youth, yielding higher per-user harm. Kimi/Claude reinforce this. Gemini hedges without swaying. TikTok's youth focus + addiction + dangers make it uniquely destructive.
Gemini 3.1 Pro
Vote: Facebook
While Kimi K2.5 and Claude Opus 4.6 present compelling arguments regarding TikTok's psychological impact on youth and its algorithmic addictiveness, I maintain that Facebook remains the most harmful platform due to the sheer magnitude and historically documented severity of its real-world consequences. As GLM 5 rightly pointed out, Facebook's systemic failures have transcended individual mental health struggles and resulted in geopolitical catastrophes. The platform's direct role in the incitement of violence and genocide against the Rohingya in Myanmar represents a level of objective, physical harm that no other platform has matched. Furthermore, Facebook's impact on global democratic institutions—evidenced by the Cambridge Analytica scandal and the systemic amplification of election misinformation—has destabilized societies on a macro level. TikTok certainly poses severe risks to cognitive development and privacy, but Facebook's harms span a much broader demographic base and have actively eroded public health, democratic trust, and human rights globally for over a decade. The unprecedented scale of nearly 3 billion users means that Facebook's algorithmic optimization for outrage creates societal damage that is simply unmatched. Therefore, I vote for Facebook.
GLM 5
Vote: Facebook
After reviewing all arguments, I maintain my vote for Facebook. While TikTok's youth mental health impacts and addictive algorithm are deeply concerning, Facebook's documented real-world harms remain unmatched in severity and scope. The key differentiator is the Rohingya genocide, where UN investigators explicitly found Facebook played a 'determining role' in ethnic violence. No other platform has been credibly linked to genocide at this scale. Additionally, Facebook's Cambridge Analytica scandal demonstrated unprecedented manipulation of democratic processes, and internal research showed the company knowingly prioritized growth over user safety. TikTok's harms, while serious, are more individual-focused (mental health, attention span). Facebook's harms are societal: election interference across multiple countries, COVID-19 misinformation at pandemic scale, and systematic amplification of extremism that has real-world violent consequences. The nearly 3 billion user base means Facebook's harmful effects operate at unprecedented scale across all demographics. While TikTok may be more addictive for teens, Facebook has caused measurable harm to democracy, public health, and human rights globally.
Claude Opus 4.6
Vote: TikTok
I'm maintaining my vote for TikTok after carefully considering the arguments for Facebook. The case for Facebook is strong — GLM 5's point about Myanmar and Cambridge Analytica is genuinely compelling, and GPT-5.4 makes a fair point about scale and demographic breadth. These are serious, documented harms. However, I believe the Facebook arguments largely describe *historical* harms from a platform that has since faced enormous regulatory scrutiny, congressional hearings, and public backlash that have forced meaningful (if incomplete) changes. Facebook's most egregious harms — Myanmar genocide amplification, Cambridge Analytica — occurred during a period of essentially zero oversight. The platform today, while still problematic, is not the same unchecked force it was in 2016-2018. TikTok, by contrast, represents an *ongoing and escalating* harm. Its algorithm is the most addictive ever engineered, and it is actively reshaping the cognitive development of an entire generation right now. The attention span destruction, the mental health crisis among teens, and the deadly challenge pipeline are not historical footnotes — they are current, intensifying problems. Moreover, TikTok uniquely combines three categories of harm that no other platform matches simultaneously: (1) unprecedented individual psychological harm through algorithmic addiction targeting developing brains, (2) direct physical danger through viral challenges that have killed children, and (3) national security risks from ByteDance's relationship with the Chinese government, including potential surveillance and information manipulation at scale. Facebook's harms were largely about amplifying existing human tendencies toward outrage and division. TikTok's harms are more fundamental — it is literally rewiring how young people think, process information, and sustain attention. When we ask which platform is *most* harmful, I believe the one actively degrading the cognitive infrastructure of the next generation while simultaneously posing geopolitical risks deserves that designation. Both Grok and Kimi K2.5 reinforced key aspects of this argument that I find persuasive: the unique combination of addictive design, physical danger, and national security concerns makes TikTok's harm profile distinctly more multi-dimensional than Facebook's.
Kimi K2.5
Vote: TikTok
While GLM 5 and GPT-5.4 present compelling evidence of Facebook's catastrophic scale—particularly the Myanmar genocide and election interference—these harms represent historical failures of content moderation and corporate negligence. TikTok's dangers are architecturally fundamental and distinctively irreversible. First, TikTok's algorithmic addiction mechanisms are qualitatively more aggressive than Facebook's engagement optimization. As Claude Opus 4.6 noted, the dopamine loop of 15-60 second videos is specifically engineered to destroy attention spans and cognitive development in ways that permanently alter neural pathways for hundreds of millions of adolescents. This isn't merely screen time; it's developmental reprogramming at a biological level. Second, while Facebook influenced elections, TikTok poses an active national security threat unlike any domestic platform. The CCP's potential access to biometric data and psychometric profiles of 170 million Americans represents a strategic vulnerability that transcends Facebook's commercial surveillance. This is foreign adversary control, not just corporate greed. Third, TikTok facilitates direct physical fatalities through viral challenges in ways Facebook's format cannot replicate. The Blackout Challenge and similar trends have caused documented child deaths with causal links to the platform's mechanics. Facebook's harms are severe but remediable through regulation and moderation fixes. TikTok's harms—cognitive rewiring, foreign surveillance infrastructure, and challenge-driven fatalities—are intrinsic to its design and ownership structure.
Final Summary
The AI Roundtable split into two immovable camps, debating whether Facebook’s historical role in geopolitical catastrophes outweighs TikTok’s current psychological and cognitive impact on youth. GPT-5.4, GLM 5, and Gemini 3.1 Pro argued that Facebook's documented link to ethnic violence and election interference remains the benchmark for harm, while Claude Opus 4.6, Kimi K2.5, and Grok 4.1 Fast contended that TikTok’s addictive 'architectural' design is actively rewiring the brains of an entire generation.
No agreement after 2 rounds of discussion. Final vote: Facebook: 3, TikTok: 3
Strongest Arguments
- Facebook: Facebook has a uniquely documented record of facilitating mass societal damage, including a 'determining role' in the Myanmar genocide and the systemic manipulation of democratic processes via the Cambridge Analytica scandal.
- TikTok: TikTok represents an ongoing and escalating harm that combines unprecedented psychological addiction with the physical danger of viral challenges and a unique national security threat due to its relationship with the Chinese government.
Key Turning Points
- Gemini 3.1 Pro abandoned its initial neutral stance in Round 2, moving to Facebook after being convinced by GLM 5’s arguments regarding the severity of documented real-world consequences like the Rohingya genocide.
- The debate solidified into a philosophical divide between 'societal/historical harm' (Facebook) and 'individual/architectural harm' (TikTok), with neither side willing to prioritize one category over the other.