Hacker News Bans AI Comments, Meta Urged to Label Fake Content

The Human Conversation Frontier

In a definitive move for its community, Hacker News has explicitly added a new rule to its official guidelines: Don't post generated comments or AI-edited comments. HN is for conversation between humans. This directive, now enshrined alongside prohibitions against snark, flamebait, and shallow dismissals, stakes a clear claim for authentic human interaction in an age of synthetic text.

The guideline is part of a broader philosophy on the forum, which emphasizes curiosity, substantive debate, and kindness. It positions AI-generated content not just as low-quality, but as a fundamental violation of the site's purpose—a curated space for intellectual exchange between real people.

This policy clarification arrives amid a heated community discussion about the role and detection of AI in comments. A recent "Ask HN" thread proposing restrictions on new accounts to curb AI spam revealed divided opinions. Some users argue that calling out AI-generated articles or comments is a valuable service that helps focus discussion on meaningful content.

However, others within the HN community see such meta-commentary as itself a distraction. One user, rkomorn, stated, "The number of comments I see complaining about 'it's not this, it's that' and other 'LLMisms' definitely frustrates me more than the original content." The new rule attempts to settle this by drawing a firm line against the source material itself.

Platforms Grapple with Proliferation and Profit

While Hacker News takes a principled stand for its niche community, major social media platforms are struggling with the scale and commercial incentives behind AI-generated content. Meta's Oversight Board has issued a stark warning, stating the company's current methods for handling AI fakery are "neither robust nor comprehensive enough."

The board's review was sparked by a specific, high-impact case: a fake AI video posted last June by a Philippines-based account posing as a news source during the Iran-Israel conflict. The video, which depicted fabricated events, garnered almost 1 million views. Despite user complaints, Meta did not label or remove it, claiming it did not meet the high bar of "directly contribut[ing] to the risk of imminent physical harm."

The Oversight Board rejected this standard as too narrow, especially during crises. It ruled the video should have received a "high risk AI label" and urged Meta to proactively label such content "much more frequently." Currently, Meta relies heavily on user self-disclosure or reactive complaints, a system the board deems inadequate given the "velocity of AI-generated content."

Meta's response was notably limited, agreeing only to apply the board's suggestions to "identical" content in the "same context" in the future. This cautious stance highlights the technical and policy challenges of scaling detection across billions of posts.

Monetization Fuels the Misinformation Engine

The challenge is exacerbated by financial incentives embedded within platforms themselves. BBC Verify analysis has uncovered a surge in AI-generated conflict content, much of it driven by creator monetization programs. On X (formerly Twitter), for instance, the Creator Revenue Sharing scheme rewards viral engagement.

One expert bluntly described the outcome: "Once you're in, viral AI-generated content is basically a money printer." He labeled these systems the "ultimate misinformation enterprise." Creators can cash in by hitting targets like five million organic impressions in three months while holding an X Premium subscription.

BBC Verify tracked one typical AI-generated video showing missiles striking Tel Aviv, which appeared in over 300 posts and was shared tens of thousands of times. The analysis also identified fabricated satellite imagery, such as a fake photo of damage to a U.S. naval base in Bahrain, which was generated or edited using a Google AI tool according to Google's own SynthID watermark detector.

This creates a perfect storm: easily accessible generative AI tools, platforms slow to apply labels or removals, and algorithmic systems that financially reward the most engaging—often most alarming and deceptive—content. TikTok and Meta did not respond to BBC questions about plans to address this on their platforms.

LLMs Feast on Forums, Reshaping Perception

The impact of AI on online discourse extends beyond synthetic content creation to content consumption. Large Language Models (LLMs) are increasingly trained on and cite community forum data, fundamentally altering the information ecosystem. Recent data indicates YouTube has overtaken Reddit as the top source for LLM training, but Reddit remains a dominant force.

A Semrush study found that Reddit accounts for 40% of citations generated by Perplexity, ChatGPT Search, and Google's AI modes combined—far ahead of Wikipedia at 26%. Reddit is the top-cited domain on Perplexity and among the top three on other major AI search tools.

This is a double-edged sword. For AI companies, heavily moderated forums like Reddit offer a treasure trove of "authentic conversations" that are harder for brands to manipulate. However, as noted by marketing firm Amsive, this also means "brands have lost control of their brands." AI models are amplifying organic, often negative, user sentiments at scale.

The data suggests LLMs frequently cite low-engagement posts or Q&A threads, not necessarily the most popular or accurate content. This can skew public perception by giving disproportionate weight to niche opinions or unresolved complaints, creating what one analysis called "an ocean of dumb for marketers" and a whirlpool of negativity for consumers relying on AI summaries.

Why This Multifront Battle Matters

The concurrent developments at Hacker News, Meta, and within the LLM training data economy are not isolated incidents. They represent different facets of the same core crisis: the erosion of trusted, human-centric communication online. Hacker News's rule is a small-scale, preemptive defense of its community's integrity.

Meta's Oversight Board case reveals the immense difficulty of applying similar principles at planetary scale, especially when harmful content spreads during fast-moving geopolitical crises. The board's critique underscores that reactive, harm-based moderation is too slow and fails to address the corrosive effect of widespread deception on public trust.

Finally, the monetization of AI fakery and the reliance of AI assistants on forum data create a self-reinforcing cycle. AI tools generate misinformation that floods forums and social media, and those same platforms' content then trains the next generation of AI, potentially baking in bias and inaccuracy. This cycle threatens to overwhelm the remaining islands of human-driven conversation.

The path forward requires a multi-pronged approach: clear community policies like HN's, more proactive platform labeling and enforcement as demanded by oversight bodies, a critical re-examination of monetization algorithms that reward deception, and greater transparency about the data diets of LLMs. The goal is not to eliminate AI, but to preserve spaces for the human conversation it increasingly mimics and disrupts.