Proactive Hybrid Recommendation Video Advertising in Generative AI Conversations: An AI-Initiated Video Advertising Format Integrating Organic and Sponsored Recommendations

This paper proposes Proactive Hybrid Recommendation Video Advertising (PHRVA), a conceptual advertising framework specifically designed for video-based interactions with generative artificial intelligence (AI). PHRVA enables a generative AI system to autonomously initiate and present a unified video recommendation sequence when the conversational context indicates a potentially appropriate opportunity, without requiring the user to explicitly request product recommendations, service suggestions, or advertisements in the current interaction. The proposed format integrates two distinct categories of recommendations into a coherent video presentation: organic video recommendations, selected or generated according to contextual relevance and potential usefulness without being included as a result of paid placement, and sponsored video recommendations, included through identifiable commercial arrangements and explicitly disclosed as sponsored content. The defining characteristic of PHRVA is that the entire AI-initiated hybrid video recommendation sequence constitutes one distinct advertising unit, rather than merely inserting an isolated video advertisement into an ordinary AI interaction or presenting sponsored videos in response to an explicit recommendation request. The system may determine an appropriate moment for initiating the sequence using authorized contextual signals derived from the current conversation, interaction history, and user-provided preferences. The resulting presentation may combine video clips, demonstrations, visual explanations, narration, subtitles, and other audiovisual elements while preserving the distinction between organic and sponsored recommendations. The framework introduces a proactive opportunity assessment mechanism, dual-source recommendation selection, hybrid video sequence construction, audiovisual disclosure, playback and interruption management, and user-control safeguards. It further establishes research propositions and an evaluation framework addressing contextual relevance, sponsorship recognition, perceived trust, audiovisual comprehension, conversational continuity, and advertising intrusiveness. This paper does not claim experimentally demonstrated effectiveness. Instead, it establishes a testable conceptual architecture for a specific form of proactive advertising in generative AI conversations, defining the complete hybrid video recommendation sequence as the advertising unit and identifying the technical, behavioral, commercial, and ethical requirements for future implementation.