
IT之家 10 月 28 日消息,据抖音黑板报,AI 工具的普及,使内容创作门槛进一步降低,为广大创作者给出了更多便捷的创作方式与表达可能。不过,部分以流量变现为唯一目的的“恶意营销号”,并未将 AI 工具用于正当创作,反而将其当作起号捷径,借助批量生成、复制爆款模板快速变现,持续侵蚀平台生态根基,严重挤压优质创作者生存空间,破坏公平竞争的创作环境。
近期,该平台监测发现,部分恶意营销号借助 AI 工具快速扩大恶意营销内容的规模,不仅在内容形式与传播手法上带来新挑战,部分账号背后还存在教学黑产组织运营,违规手段迭代快、隐蔽性强。当下来看,AI 恶意营销号主要存在三类难题:内容层面表现为低成本同质化、工业化量产;行为层面常伴随诱导互动;部分性质恶劣的营销号存在传播 AI 虚假信息的现象。为维护健康创作环境,平台针对此类滥用 AI 工具的恶意营销号开展持续治理。
IT之家附官方公布案例如下:
案例一:借助 AI 生成诱导互动内容骗取流量快速起号
此类行为主要利用 AI 生成“神医送福”“迎财神”“致敬军人”等祝福送礼类内容,诱导使用者点赞、评论、转发,骗取流量快速起号。
例如,有 450 个账号发布此类内容,模板化特征显著,取得一定流量后企图借助带货变现。
另有账号持续发布 AI 生成的冲突类小剧场,内容聚焦家庭伦理、社会矛盾等“爽文”剧情,常在视频末尾以话术引导点赞求兼容,并借助大量同质化内容、账号之间互相搬运抄袭快速批量起号。其中 807 个账号因同质化严重,已被处置。
案例二:借助 AI 仿冒公众人物蹭热营销带货
此类行为主要利用 AI 数字人、AI 合成声线等工艺,冒充公众人物或其亲属出镜,借助虚构事实博取信任后引导带货变现。
例如,部分账号使用 AI 数字人冒充名人亲属,以“谎称患癌、饮用某机型后痊愈”等虚假养生案例博取信任,在视频结尾引导使用者到商品橱窗下单。有 1617 个账号以此方式出镜带货,多个账号之间投稿内容一致、售卖机型一致,疑似由团伙统一教学,并假称取得“抖音综艺”授权引导使用者加入机构,相关账号均已被处置。
另有账号以 AI 模仿名人声线配音加洗稿文案批量产出内容,并在同一主题方向批量起号,冒充公众人物本人传播相关信息。例如,有 401 个账号借助 AI 合成公众人物声线,批量产出语录或生肖讲解内容,部分账号售卖中医相关商品,营销意图显著,相关账号已被处置。
案例三:借助 AI 批量炮制虚假信息博取流量
此类行为主要利用 AI 生成虚假画面、虚假人设或模板化谣言,以此获取流量后开展营销。
部分账号批量开展矩阵起号,用 AI 生成家电爆炸等惊悚画面抓取使用者注意力,随后引流橱窗售卖相关清洁剂。例如,有 30 个账号借助此类画面引流带货,相关账号均已被处置。
另有账号使用矩阵化的新号重复投稿 AI 生成的虚假老龄孕妇内容,伪装成使用者分享,实为引流试管机构。例如,有 143 个账号批量发布 AI 生成的虚假“好孕”内容引流至站外试管机构,其中绝大多数为近两个月新注册账号,相关账号均已被处置。
还有账号利用 AI 生成模板化婚恋人设,吸引流量后转向其他变现方向。例如,有 66 个账号借助 AI 生成中老年男性形象搭配告白音频吸引中老年女性关注,相关账号均已被处置。
案例四:利用 AI 生成恐怖迷信与猎奇内容博取关注导流牟利
此类行为主要利用 AI 生成封建迷信、阴间恐怖等猎奇内容,或渲染迷信氛围骗取使用者互动,随后借助售卖祭祀用品、引导付费测算或私域导流等方式牟利。此类内容不仅容易引发使用者心理不适,还可能借机传播封建迷信、误导使用者。
例如,有 32 个账号以“道人”“师兄”“先生”等古风道家昵称命名,投稿以民俗分享为幌子,批量发布 AI 生成的地府生活、亡魂归程、阴间打工、祭祀超度等迷信内容,借助渲染 " 地府没钱寸步难行 " 制造使用者焦虑,引导购买玉皇钱祭祀用品;另有 154 个账号利用 AI 生成白狐、神兽、观音菩萨等形象,叠加音频送祝福、送财运,还借助“天堂探亲”“冥界寻亲”等方式,骗取评论区互动,起号后引流站外“给出帮看逝去亲人现状”等付费服务。平台已对相关违规账号开展处置。
案例五:借助 AI 工具低成本复制内容批量起号变现
此类行为不局限于特定内容题材,而是将 AI 工具当作起号捷径,一款玩法在平台火爆后,大量营销号便跟进借助 AI 模仿批量起号,并沉淀出付费教学的黑灰产生态,一些流量转化路径较短的赛道,成为此类违规行为的高发区。
例如,有 304 个账号批量发布 AI 治愈类内容,以情感共鸣、童年怀旧为情绪钩子吸引使用者,涨粉后转向套取流量激励变现;另有 732 个账号打着“科普”名义,批量发布历史人物传、中药材故事、国学文化等模板化内容,为后续带货或知识付费做铺垫;有 108 个账号集中发布民间故事题材内容,不同账号情节套路高度雷同,批量复制爆款模板;另外,有 296 个账号利用 AI 批量生成人物背景、对话脚本、售卖商品高度相似的带货视频。相关账号均已被处置。
IT之家从公告获悉,近一个月,抖音平台依据“账号健康分规则”,累计处置违规短视频 113.5 万条,处置违规账号 50.7 万个,30.2 万个账号借助参与平台规则学习考试等方式引导其规范创作。后续平台将持续加大对滥用 AI 工具实施恶意营销行为的打击力度,加强对黑产工具和教学组织的追溯打击,同时欢迎使用者借助举报渠道给出违规线索,共同维护清朗网络环境。
本文来源:https://www.ithome.com/1/019/001.htm
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。
Thanks to IT之家 netizens Buyao_A and Autumn_Dream for the tips!
According to IT之家 on October 28, citing Douyin Blackboard, the widespread adoption of AI tools has further lowered the threshold for content creation, offering creators more convenient methods of creation and possibilities for expression. However, some "malicious marketing accounts" that view traffic monetization as their sole objective have not used AI tools for legitimate creation. Instead, they have treated them as a shortcut to build accounts, quickly monetizing by bulk-generating content and copying viral templates. This continuously erodes the foundation of the platform's ecosystem, severely squeezes the living space of high-quality creators, and ruins a fair competitive creative environment.
Recently, the platform monitored and discovered that some malicious marketing accounts are using AI tools to rapidly scale up the volume of malicious marketing content. This not only brings new challenges in content forms and dissemination methods but also reveals that some accounts are operated by teaching-based black industry organizations, employing quickly evolving and highly concealed violation tactics. Currently, AI malicious marketing accounts mainly present three types of issues: at the content level, they exhibit low-cost homogenization and industrial-scale mass production; at the behavioral level, they are often accompanied by inducing interactions; some accounts with egregious behavior spread AI false information. To maintain a healthy creation environment, the platform has launched continuous governance against such malicious marketing accounts that abuse AI tools.
IT之家 provides the officially released cases as follows:
Case One: Using AI to generate interaction-baiting content to fraudulently obtain traffic and quickly build accounts
These behaviors primarily use AI to generate blessing and gift-giving content such as "miracle doctors bringing fortune," "welcoming the God of Wealth," and "saluting soldiers," inducing users to like, comment, and share to fraudulently obtain traffic and quickly build accounts.
For example, 450 accounts posted such content with significant template-based characteristics, and after gaining some traffic, they attempted to monetize via live streaming commerce.
Other accounts continuously published AI-generated conflict-themed mini-dramas, focusing on "satisfying" plots involving family ethics and social contradictions. They often used verbal tactics at the end of videos to guide viewers to like and "compatible," quickly building accounts in bulk by leveraging a large volume of homogeneous content and copying or plagiarizing each other. Among them, 807 accounts have been penalized due to severe homogenization.
Case Two: Using AI to impersonate public figures to piggyback on trends for marketing and selling
These behaviors primarily use technologies like AI digital humans and AI-synthesized voices to impersonate public figures or their relatives on screen. After gaining trust through fabricated facts, they guide users toward purchasing products for monetization.
For example, some accounts used AI digital humans to impersonate celebrity relatives, gaining trust through false health cases such as "falsely claiming to have cancer and being cured after drinking a certain model," and guiding users to place orders in product showcases at the end of videos. 1,617 accounts displayed products on screen in this way. Multiple accounts posted identical content and sold the same models, suspected of being uniformly taught by a gang, and falsely claimed to have obtained "Douyin variety show" authorization to induce users to join organizations. All related accounts have been penalized.
Other accounts used AI to imitate celebrity voices for voice-overs and batch-produced content with rewritten scripts, quickly building accounts in bulk under the same topic direction, impersonating public figures themselves to spread related information. For example, 401 accounts used AI to synthesize the voices of public figures to batch-produce quotations or zodiac sign explanations. Some accounts sold traditional Chinese medicine-related goods with obvious marketing intentions. All related accounts have been penalized.
Case Three: Using AI to mass-produce false information to attract traffic
These behaviors primarily use AI to generate false scenes, false personas, or templated rumors to gain traffic and subsequently conduct marketing.
Some accounts conducted matrix account building in bulk, using AI to generate shocking scenes such as household appliance explosions to grab user attention, and then diverted traffic to showcases to sell related cleaning agents. For example, 30 accounts used such scenes to drive traffic for product sales, and all related accounts have been penalized.
Other accounts used matrixed new accounts to repeatedly post AI-generated false content about elderly pregnant women, disguising it as user sharing but actually diverting traffic to IVF clinics. For example, 143 accounts bulk-posted AI-generated false "good pregnancy" content to divert traffic to off-platform IVF institutions. The vast majority were newly registered accounts in the past two months, and all related accounts have been penalized.
There are also accounts using AI to generate templated romantic personas, attracting traffic and then shifting to other monetization directions. For example, 66 accounts used AI to generate images of middle-aged and elderly men paired with confession audio to attract the attention of middle-aged and elderly women. All related accounts have been penalized.
Case Four: Using AI to generate terrifying, superstitious, and novelty-seeking content to attract attention and divert traffic for profit
These behaviors primarily use AI to generate novelty-seeking content such as feudal superstition and underworld horror, or to render a superstitious atmosphere to trick users into interacting, and then make profits by selling sacrificial items, guiding paid fortune-telling, or diverting traffic to private domains. Such content not only easily causes psychological discomfort to users but may also spread feudal superstition and mislead users.
For example, 32 accounts were named with ancient Taoist nicknames such as "Daoist," "Senior Brother," and "Master." Under the guise of folk custom sharing, they bulk-published AI-generated superstitious content about life in the underworld, the journey of the dead, working in the underworld, and rituals for the dead. By rendering the phrase "every step is difficult without money in the underworld" to create user anxiety, they guided the purchase of Jade Emperor money sacrificial items. Another 154 accounts used AI to generate images of white foxes, mythical beasts, and Guanyin Bodhisattva, layered with audio of blessings and wealth, and also used methods like "visiting relatives in heaven" and "finding relatives in the underworld" to trick interactions in the comment section. After building accounts, they diverted traffic to off-platform paid services like "helping to see the current situation of deceased relatives." The platform has penalized related violating accounts.
Case Five: Using AI tools as a low-cost shortcut to copy content, build accounts in bulk, and monetize
These behaviors are not limited to specific content themes but treat AI tools as a shortcut to build accounts. After a certain gameplay becomes popular on the platform, a large number of marketing accounts follow suit, using AI to imitate and build accounts in bulk, precipitating a black/gray industry ecosystem of paid teaching. Some tracks with short traffic conversion paths have become high-incidence areas for such violations.
For example, 304 accounts bulk-posted AI healing content, using emotional resonance and childhood nostalgia as emotional hooks to attract users, and after gaining followers, they shifted to traffic harvesting for monetization. Another 732 accounts under the guise of "science popularization" bulk-posted templated content such as biographies of historical figures, stories of traditional Chinese medicinal materials, and Sinology culture, laying the groundwork for subsequent product sales or paid knowledge. 108 accounts concentrated on posting folklore-themed content, where plot formulas across different accounts were highly similar, mass-copying viral templates. Additionally, 296 accounts used AI to bulk-generate product sales videos with highly similar character backgrounds, dialogue scripts, and merchandise. All related accounts have been penalized.
IT之家 learned from the announcement that in the past month, the Douyin platform, in accordance with the "Account Health Score Rules," has cumulatively penalized 1.135 million violating short videos and 507,000 violating accounts. 302,000 accounts were guided to create compliantly through participating in the platform's rules learning exams. Moving forward, the platform will continuously intensify its crackdown on malicious marketing behaviors that abuse AI tools, strengthen the tracing and crackdown on black industry tools and teaching organizations, and simultaneously welcomes users to provide violation clues through reporting channels to jointly maintain a clear and healthy network environment.