It's Not Me [pdf] (www.berkshirehathaway.com)

Berkshire Hathaway issued a short statement warning that several YouTube videos use AI-generated images and voice impersonations to mimic Warren Buffett, noting the audio often sounds like a flat, monotone impersonator and urging viewers to be cautious. The company linked to a specific channel and said these fake clips could mislead people unfamiliar with Buffett. Berkshire also announced it will publish an official press release and message from Buffett on its website on November 10 and advised shareholders to rely on that channel for authentic communications. The alert underscores a broader AI/ML risk: increasingly accessible generative models for image synthesis and voice cloning make high-profile deepfakes easy to produce, enabling misinformation, fraud, and social-engineering attacks that can affect markets and reputations. Technically, current detection cues—unnatural prosody, visual artifacts, inconsistent lip-sync—remain useful but are being eroded as models improve. The incident highlights the need for stronger provenance and authentication (cryptographic signing, verified channels, provenance metadata), improved platform moderation, and wider deployment of robust detection tools and literacy among journalists, investors and the public to avoid being misled by synthetic media.