Misleading AI-generated doctors pose 'huge danger to public safety' 误导性AI生成医生对公众安全构成“巨大危险”

TL;DR

• AI-generated fake doctors on TikTok are spreading dangerous health misinformation, with 40% of top health videos being AI-generated or AI-assisted.
• These accounts promote debunked myths (e.g., microwaving food in plastic causes cancer) and bizarre remedies like Coca-Cola mixed with onion as pain relief.
• Top AI-generated health videos average 2.5 million views, indicating widespread public exposure to unverified medical claims.
• Medical professionals warn this poses a “huge danger to public safety,” especially as one in five people use social media for health information.
• TikTok denies the findings reflect platform reality, citing removal policies and partnerships with WHO and NHS.

Why It Matters

This issue highlights the urgent need for robust AI content moderation and public health literacy as generative AI enables scalable, convincing misinformation that can directly harm individuals’ health decisions. The scale of reach—millions of views per video—demonstrates how quickly false medical narratives can spread unchecked, undermining trust in legitimate healthcare sources and potentially leading to real-world physical harm.

Technical Details

• Research conducted by Hallam analyzed 1,198 TikTok videos across health-related search terms such as “health tips” and “doctor advice,” selecting the top 50 results per term.
• AI-generated content was identified through visual and textual cues consistent with deepfake avatars and scripted, non-clinical language patterns typical of synthetic media.
• For specific queries like “health tips,” 84% of top videos were classified as AI-generated or AI-assisted, suggesting algorithmic amplification of synthetic health content.
• Videos featured anthropomorphic AI doctors promoting pseudoscientific claims without disclaimers, often using professional-looking visuals to mimic real medical authority.
• TikTok responded by removing flagged content sent by The Guardian but maintained its stance that external research does not accurately represent platform governance efforts.

Industry Insight

Social media platforms must implement stricter detection and labeling systems for AI-generated health content, particularly when it mimics licensed professionals or makes clinical claims. Collaborations between tech companies, medical institutions, and regulators should establish real-time verification protocols to flag or suppress high-risk misinformation before it gains traction. Additionally, integrating AI literacy into public health campaigns will empower users to critically evaluate online medical advice, reducing vulnerability to exploitative synthetic content designed purely for engagement.

TL;DR

• 英国医学协会警告AI生成医生账号传播虚假医疗建议，对公众安全构成“巨大威胁”。
• Hallam机构研究发现TikTok上40%的健康类热门视频由AI生成或辅助，部分关键词下该比例高达84%。
• AI账号推广未经证实的癌症谣言（如微波炉加热塑料致癌）及虚假产品（如“Hyalethinap Plus Pro Max”），平均每条视频获250万次观看。
• NHS和皇家全科医师学院指出此类内容导致患者困惑，可能因错误建议损害健康甚至危及生命。
• TikTok否认研究准确性，称其已移除有害信息并与WHO、NHS合作提供可靠健康资讯。

为什么值得看

本文揭示了AI技术在医疗健康领域的滥用风险，尤其针对社交媒体平台中深度伪造医生的泛滥现象，对公众健康构成直接威胁。对于AI从业者而言，这凸显了内容真实性验证与平台监管机制的紧迫性；对行业而言，它警示了生成式AI在敏感领域落地时缺乏伦理约束可能引发的社会危机。

技术解析

• 数据来源与方法：Hallam机构分析1,198个视频，选取“健康提示”“医生建议”等搜索词前50条结果，识别AI生成或辅助内容占比。
• AI应用形式：使用虚拟形象（avatar）模拟真实医生发布视频，结合文本生成与语音合成技术制造可信度高的伪专业内容。
• 典型误导案例：包括将日常行为（如用塑料盒微波食物、使用除臭剂、睡前放手机）错误关联至癌症成因；推销无科学依据的抗衰老产品“Hyalethinap Plus Pro Max”。
• 传播效果量化：高热度AI健康视频平均获得250万浏览量，显示算法推荐机制加速虚假信息扩散。
• 平台响应局限：TikTok声称已部署AI标签技术与WHO/NHS合作，但被批应对速度滞后且未主动拦截批量生成的虚假账号。

行业启示

• 平台责任重构：社交网络需建立更严格的AI内容标识与审核标准，特别是在医疗、金融等高敏感领域，应引入第三方权威机构实时校验机制。
• 公众素养升级：教育体系应纳入数字健康辨别能力培养，帮助普通用户识别AI生成内容的特征（如语气过度统一、缺乏具体病例细节）。
• 技术治理前置：AI模型开发阶段即须嵌入“防滥用”模块，例如限制医疗类问答输出未经证实的治疗方案，并强制要求来源标注与风险提示。