It’s 8pm somewhere in the world and two Tubi TV subscribers are logging into their accounts. One is from Canada, the other in Australia. Each expects a personal 1-1 interaction with the Tubi AI to find something to watch. What they don’t see? That the Tubi AI is delivering that exact same personal experience to their 100 million subscribers at the same time.
Even as AI chatbots grow in popularity to help subscribers find the movies and shows they want to watch, their trust in AI chatbots hasn’t kept pace. Why? The LLMs are trained on data that goes stale fast. Ask a chatbot what to watch tonight and it might give you an answer based on information that’s months old. That’s why 50% of viewers still trust traditional search over AI, a 2 to 1 gap.
What a Tubi TV subscriber wants to watch in Mexico City is different from the subscriber in Toronto, Canada. Not only are the content options different, from licensing, language, cultural context and local events (example, Cinco de Mayo, that shapes what that Mexican subscriber wants to watch on that day), but the first party (1P) data is siloed or stale. And the AI chatbot doesn’t know what to recommend to the subscriber. The result? The same generic recommendation is given to both of them, that misses the mark entirely. That’s why the trust gap exists. Both subscribers are fact checking the AI recommendation, not because it was entirely wrong, but because it wasn’t right for them, where they are, right now.
The AI chatbot doesn’t know if you’re in Mexico City or Toronto. It doesn’t know it’s Cinco de Mayo. And right now, that gap between what the AI assumes and what you actually want is costing streaming platforms your trust. Here’s my question to you:
Do you use an AI chatbot to find something to watch? And do you trust it? Let me know in the comments.