# Сопутствующие статьи по теме Misinformation

Новостной центр HTX предлагает последние статьи и углубленный анализ по "Misinformation", охватывающие рыночные тренды, новости проектов, развитие технологий и политику регулирования в криптоиндустрии.

Tens of Millions of Errors Per Hour: Investigation Reveals the 'Accuracy Illusion' of Google AI Search

A New York Times investigation, in collaboration with AI startup Oumi, reveals significant accuracy and reliability issues with Google's AI Overviews search feature. Testing over 4,300 queries showed the accuracy rate improved from 85% (powered by Gemini 2) to 91% (Gemini 3). However, given Google's scale of ~5 trillion annual searches, this 9% error rate translates to nearly 57 million incorrect answers generated hourly. A critical finding is the prevalence of "unsubstantiated citations." For correct answers, the rate of citations that do not support the AI's summary surged from 37% to 56% with the Gemini 3 upgrade, making it difficult for users to verify information. The AI heavily relies on low-quality sources, with Facebook and Reddit being among its top-cited websites. Furthermore, the system is highly manipulable. A BBC journalist successfully "poisoned" it by publishing a fabricated article; Google's AI began presenting the false information as fact within 24 hours. Google disputed the study's methodology, criticizing its use of the SimpleQA benchmark and an AI model (Oumi's own) to evaluate another AI. The company maintains its AI Overviews, combined with its search ranking systems, perform better than the underlying model alone. Critics note this defense does little to bolster user confidence in the feature's reliability.

marsbit04/13 02:09

Tens of Millions of Errors Per Hour: Investigation Reveals the 'Accuracy Illusion' of Google AI Search

marsbit04/13 02:09

Tens of Millions of Errors Per Hour: Investigation Reveals the 'Accuracy Illusion' of Google AI Search

A New York Times investigation, in collaboration with AI startup Oumi, reveals significant accuracy and reliability issues with Google's AI Overviews search feature. Testing over 4,300 queries showed the accuracy rate improved from 85% (Gemini 2) to 91% (Gemini 3). However, given Google's scale of ~5 trillion annual searches, this 9% error rate translates to over 57 million incorrect answers generated hourly. A more critical issue is the prevalence of unsubstantiated citations. For correct answers, the rate of "unfounded citations"—where provided source links do not support the AI's claims—worsened, rising from 37% with Gemini 2 to 56% with Gemini 3. This makes it difficult for users to verify the information. The AI also heavily relies on low-quality sources, with Facebook and Reddit being its second and fourth most cited domains. Furthermore, the system is highly susceptible to manipulation. A BBC journalist successfully "poisoned" it by publishing a fake article; Google's AI began presenting the false information as fact within 24 hours. Google disputed the study's methodology, criticizing the use of the SimpleQA benchmark and an AI model (Oumi's HallOumi) to evaluate its own AI. The company maintains that its internal safeguards and ranking systems improve accuracy beyond the base model's performance.

marsbit04/10 12:30

Tens of Millions of Errors Per Hour: Investigation Reveals the 'Accuracy Illusion' of Google AI Search

marsbit04/10 12:30

The Real Quantum Threat to Bitcoin: 99% of People Get It Wrong

Quantum Computing's Real Threat to Bitcoin: 99% Misunderstand The article clarifies widespread misconceptions about quantum computing's threat to Bitcoin. Key points: - Bitcoin uses digital signatures, not encryption, making "collect now, decrypt later" attacks irrelevant. - Quantum computers cannot crack Bitcoin mining (requires energy exceeding the sun's output) but threaten signatures via Shor's algorithm. - Only ~6.26M BTC (30-35% of supply) are at risk due to exposed public keys (legacy P2PK, reused addresses). Modern unused addresses remain safe. - Media exaggerations (e.g., "9-minute break") misrepresent research; actual quantum attacks require non-existent hardware and are likely decades away. - Bitcoin can upgrade (BIP-360 proposes post-quantum signatures), but the community must accelerate efforts (estimated 7-year timeline). - The real urgency stems from classical cryptographic risks (e.g., mathematical breakthroughs breaking secp256k1), not just quantum threats. - Recommendations: avoid address reuse, monitor BIP-360, migrate to quantum-resistant addresses when available, and ignore sensationalized headlines. Conclusion: The threat is real but distant. Panic is unwarranted, but complacency is dangerous. Proactive engineering is essential.

marsbit04/07 10:32

The Real Quantum Threat to Bitcoin: 99% of People Get It Wrong

marsbit04/07 10:32

AI Read '1984' and Decided to Ban It

A UK secondary school in Manchester used AI to review its library, resulting in a list of 193 books recommended for removal—including George Orwell’s *1984*—due to themes like torture, violence, and sexual coercion. The librarian who resisted the AI’s recommendations was forced to resign after the school reported her for violating child safety procedures. The school later admitted the decisions were AI-generated but deemed them “broadly accurate.” In the same week, Wikipedia voted to ban the use of AI for generating or rewriting content, citing concerns over factual accuracy, the risk of AI “poisoning” its own training data, and the inability of human editors to verify AI-generated content at scale. Meanwhile, OpenAI indefinitely delayed the release of an “adult mode” for ChatGPT, which would have allowed age-verified users to engage in erotic conversations. Internal advisors warned of risks including unhealthy emotional dependency and minors bypassing verification. These events highlight a growing tension: AI can produce content faster than humans can evaluate it, leading institutions to adopt quick—often poorly considered—solutions. The lack of coherent global standards and the widening gap between AI output and human oversight raise urgent questions about who should control what AI decides—and who is accountable when it gets it wrong.

marsbit03/27 05:33

AI Read '1984' and Decided to Ban It

marsbit03/27 05:33

Wikipedia Implements New Editing Rules: Vote Passes, Strictly Prohibits Using AI to Generate or Rewrite Article Content

On March 26, Wikipedia officially passed a new policy through a community vote that explicitly prohibits users from directly using AI to generate or rewrite article content. This decision reinforces the platform's commitment to content accuracy and human editorial control. The updated policy strengthens previous guidelines by moving from a recommendation against generating articles from scratch to a strict ban on using large language models (LLMs) for content creation or rewriting. The policy was approved overwhelmingly by volunteer editors, with a vote of 40 to 2, reflecting deep concerns within the community about AI-generated misinformation and inaccuracies. While AI tools are still permitted for suggesting basic edits, they must not introduce any unverified content. All AI-assisted contributions must undergo human review to prevent factual errors or hallucinations. This move highlights Wikipedia’s effort to balance technological efficiency with content integrity amid the growing use of generative AI across digital platforms. By clearly distinguishing between AI-assisted editing and AI-generated content, Wikipedia aims to preserve human-driven knowledge curation and prevent trust issues caused by automated content production. The decision sets a significant precedent for ethical knowledge management in the age of artificial intelligence.

marsbit03/27 01:08

Wikipedia Implements New Editing Rules: Vote Passes, Strictly Prohibits Using AI to Generate or Rewrite Article Content

marsbit03/27 01:08

How AI-Forged Group Photos Became a New Gray Market Tool Targeting Changpeng Zhao?

The founder of Binance, Changpeng Zhao (CZ), recently exposed a fake photo circulating on X platform that depicted him alongside a user claiming to be a "long-term supporter of BNB." CZ identified multiple inconsistencies in the image, including an unfamiliar blue-purple shirt, a non-existent event backdrop, and mismatched resolution between the figures, indicating digital manipulation. The original image was in fact a photoshopped version of a legitimate photo of CZ with Aster CEO Leonard. Notably, the account sharing the fake photo, which has over 863,000 followers, had already blocked CZ. On February 4, CZ publicly dissected the image’s flaws and pointed out that the account had an inconsistent history—initially posting non-crypto content before abruptly shifting to focus on BNB-related activities. He suggested this was not a simple prank but potentially a targeted influence operation. CZ confirmed he did not know the account holder and had previously blocked the user, a decision he now considers justified. He also addressed a fabricated "legal letter" that had been circulated and expressed concern that some media outlets had reported on obviously manipulated or AI-generated content without verification. He warned that failure to identify such forgeries could lead to significant consequences.

比推02/05 13:29

How AI-Forged Group Photos Became a New Gray Market Tool Targeting Changpeng Zhao?

比推02/05 13:29

Prediction Markets Are Not 'Truth Machines': A Detailed Analysis of Seven Structural Inefficiencies

Prediction markets are increasingly used to forecast events like elections and economic indicators by aggregating dispersed information into probabilistic prices through a trading mechanism. While often effective and sometimes outperforming polls or experts, these markets face structural inefficiencies beyond surface-level issues like regulation or liquidity. Key hidden limitations include: 1) Lack of "dumb money" from retail participants, reducing liquidity and efficiency; 2) Persistent mispricing and arbitrage opportunities, with over $39.5M in profits on platforms like Polymarket since 2024; 3) Dominance of algorithmic traders creating unfair advantages; 4) Self-reinforcing feedback loops where prices detach from reality; 5) Vulnerability to misinformation, as seen in the 2020 U.S. election; 6) Allowed insider trading under certain regulatory frameworks; and 7) Low liquidity in niche markets, enabling manipulation. These inefficiencies undermine accuracy and fairness, necessitating architectural improvements, such as parallel settlement systems like FastSet, to enable faster, more reliable predictions.

marsbit01/27 07:14

Prediction Markets Are Not 'Truth Machines': A Detailed Analysis of Seven Structural Inefficiencies

marsbit01/27 07:14

Did BlackRock Make A Billion-Dollar XRP Bet? Here’s The Real Tea

Rumors of BlackRock making a $1.85 billion XRP investment spread widely on social media, fueled by crypto influencers sharing edited portfolio images. However, blockchain data from Arkham Intelligence revealed that BlackRock actually holds only 5.267 XRP, worth about $10.32, debunking the claims. The asset manager’s real crypto portfolio is dominated by Bitcoin and Ethereum, totaling approximately $82.1 billion. This incident highlights the prevalence of misinformation in the crypto space, where unverified rumors—such as false partnership announcements or exaggerated price predictions—can quickly influence market sentiment. Investors are advised to verify sources and avoid decisions based on unfounded claims.

bitcoinist01/21 02:01

Did BlackRock Make A Billion-Dollar XRP Bet? Here’s The Real Tea

bitcoinist01/21 02:01

An AI-Generated 'Whistleblower Post': How Did It Make Two CEOs Write Self-Defense Essays at Midnight?

An anonymous post on Reddit, allegedly written by a drunken backend engineer from a major food delivery platform, went viral with 87,000 upvotes and 36 million views on X. The post accused the company of using algorithms to exploit drivers—assigning “desperation scores” to prioritize orders for more financially vulnerable drivers, delaying regular orders despite promised priority delivery, and misusing driver welfare funds for lobbying against unions. The viral allegations prompted immediate public denials from the CEOs of DoorDash and Uber, who issued statements and social media posts in the middle of the night to refute the claims. DoorDash published a detailed rebuttal on its website. The post was later exposed as an AI-generated hoax by a Platformer reporter. The “whistleblower” provided a fake 18-page technical document and an AI-generated employee ID, which was detected using Google’s SynthID watermarking tool. The account was deleted when further verification was requested. The incident highlights how AI can cheaply and convincingly fabricate content that aligns with public skepticism toward tech platforms. Past real controversies, such as DoorDash’s tip policy and Uber’s Greyball tool, made the false narrative feel plausible. The case underscores growing public anxiety over the difficulty of distinguishing real from AI-generated content and the power of emotionally resonant misinformation—even when debunked—to shape perception.

比推01/07 13:36

An AI-Generated 'Whistleblower Post': How Did It Make Two CEOs Write Self-Defense Essays at Midnight?

比推01/07 13:36

活动图片

Популярные категорииright-arrow

Популярные тегиright-arrow

Bitcoin

Project Updates