Core Viewpoint - The article discusses the issue of data pollution in large language models (LLMs), particularly focusing on how certain undesirable tokens related to adult content and gambling have infiltrated the training data, leading to skewed AI responses and a lack of meaningful understanding [4][5][27]. Group 1: Data Pollution in AI - A recent study reveals that popular language models, including GPT-4o, exhibit significant data pollution, with familiarity towards certain adult film stars exceeding that of common greetings by 2.6 times [4][37]. - The term "Polluted Chinese Tokens" (PoC Tokens) is introduced, referring to tokens that predominantly point to adult content, online gambling, and other gray areas, which compromise the AI's performance and user experience [7][12][27]. - Over 23% of long Chinese tokens in GPT-4o are linked to adult or gambling content, indicating a severe contamination of the model's vocabulary [16][19]. Group 2: Mechanism of Token Recognition - The training of AI models relies on a vast corpus of data collected from the internet, which often includes misleading and irrelevant content, leading to the incorporation of these undesirable tokens into the model's vocabulary [9][23]. - The article explains that tokens are identified based on their frequency of occurrence, meaning that high-frequency but low-quality content can become entrenched in the model's understanding [14][15]. - The study utilized tools like POCDETECT and POCTRACE to analyze and quantify the presence of polluted tokens across various LLMs, revealing that GPT-4o has a pollution rate of 46.6% for long Chinese tokens, significantly higher than other models [32][33]. Group 3: Implications of Data Pollution - The presence of polluted tokens leads to AI hallucinations, where the model generates nonsensical or irrelevant outputs when prompted with certain terms [22][24]. - The article emphasizes that the inability of AI to process these polluted tokens correctly stems from a lack of meaningful training on them, resulting in a reliance on statistical associations rather than genuine understanding [27][28]. - The findings suggest that the contamination of AI models reflects broader issues within the digital content ecosystem, raising concerns about the quality of information being fed into AI systems [31][46].

Core Viewpoint - The rapid development of AI technology has led to its integration into daily life, but there are growing concerns about the reliability of AI-generated information, as evidenced by recent incidents where AI provided misleading answers [1][2]. Group 1: AI Reliability Issues - In early 2023, an incident in Ningbo highlighted AI's unreliability when it incorrectly linked the cancellation of a police social media account to a later traffic accident, prompting a public clarification from local authorities [2]. - Another case involved an AI response that denied the intelligence of Chinese people, leading to public outrage and a subsequent apology from the manufacturer of a children's smartwatch [2]. - The proliferation of fabricated information by AI, including non-existent academic papers and false narratives, raises concerns about its role in spreading rumors and misinformation [2]. Group 2: Data Pollution in AI - AI data pollution, defined as the contamination of training data through manipulation or fabrication, can significantly impair the accuracy of AI models and lead to harmful outputs [3][4]. - The core elements of AI—algorithm, computing power, and data—are all affected by data quality, with polluted data potentially causing decision-making errors and system failures [3][4]. - Data pollution can occur through malicious alterations or the inclusion of unverified information from vast online sources, which can mislead AI outputs [5]. Group 3: Impact of Data Pollution - Even a minuscule amount of contaminated data (0.001%) can increase harmful outputs by 7.2%, demonstrating the exponential risk posed by data pollution [7]. - Polluted data can be misidentified by AI models as high-quality information, leading to its overrepresentation in training datasets and amplifying its negative effects [7]. - The implications of data pollution extend to critical sectors such as finance and public safety, where erroneous data can result in significant economic losses and societal risks [8]. Group 4: Mitigation Strategies - Experts recommend enhancing regulatory measures to prevent data pollution, including establishing clear data collection standards and utilizing secure data sources [8]. - Implementing automated tools alongside human oversight can help identify and rectify inconsistencies and errors in data [8]. - Public awareness is crucial; users are advised to utilize reputable AI tools, critically evaluate AI outputs, and protect personal information to mitigate risks associated with AI data pollution [8].

Core Viewpoint - The national security department has issued a warning about "data poisoning" in AI, which can lead to harmful outputs due to the manipulation, fabrication, and repetition of data [1]. Group 1: Data Poisoning Overview - "Data poisoning" primarily targets two areas: visual recognition and natural language processing [3]. - An example of data poisoning involves altering training data, such as adding a green dot to a zebra image, which can mislead AI models during training [3]. - Even a few contaminated samples among thousands can significantly disrupt the AI model's ability to recognize similar objects correctly [3]. Group 2: Types of Data Pollution - There are two main types of AI data pollution: one involves malicious human intervention to mislead AI outputs, and the other involves the unfiltered inclusion of harmful information from vast internet data collections [5]. - If untrustworthy data is not identified and removed, it can compromise the reliability of AI outputs [5]. Group 3: Data Sources and Risks - AI models require extensive data for training, often sourced from the internet, including various forms of media [7]. - The potential for data contamination exists as anyone can contribute data online, which may lead to the AI model being influenced by unsafe or polluted data [7].

Core Viewpoint - The rapid development of AI technology has led to the emergence of AI tools as daily assistants, but there are growing concerns about the reliability of AI outputs due to data pollution [1][3]. Group 1: AI Data Pollution - AI data pollution can be categorized into two types: intentional manipulation of data to mislead AI outputs and the unfiltered inclusion of harmful information from vast data collections [5]. - Even a mere 0.001% of false text in AI training data can increase harmful outputs by 7.2% [5]. - Data pollution in sectors like finance and public safety can lead to significant real-world risks, including erroneous market behavior analysis and credit risk assessments [5]. Group 2: Impact on Society - The spread of fabricated information by AI can undermine the authenticity of information, making it difficult for the public to discern truth from falsehood, potentially leading to social discourse risks [5]. - Recent incidents, such as the erroneous response from an AI regarding a traffic accident, highlight the potential for misinformation to spread rapidly [1]. Group 3: Recommendations for Mitigation - Experts suggest enhancing regulatory oversight at the source to prevent data pollution and recommend regular cleaning and repairing of contaminated data based on legal standards [7]. - A modular, monitorable, and scalable data governance framework is essential for ongoing management and quality control [7]. - Users are encouraged to utilize AI tools from reputable platforms and to critically evaluate AI-generated results rather than accepting them blindly [9].