Hands-on with Amazon’s new “AI-enhanced” Fire TV search

Navigating the sheer volume of streaming content available today can be a full-time job. Recommendations from friends, blog posts, and those TikToks of movies I haven’t thought about in years all help. But finding something myself, my husband, my 13-year-old daughter, and my 16-year-old son all want to watch together is still a herculean task.

So, when Amazon announced its new AI-powered voice search function for Fire TVs at its fall event last year, I was intrigued. With its promise to make searching for content easier and smarter, I hoped it would be the solution to my problems. I’ve now had some hands-on time with the new feature, and while it shows promise, like a lot of AI-powered search right now, it’s just not reliable enough to be all that useful.

The basic idea is that you can use more natural language to ask Alexa to find you something to watch. Whether you have a show in mind but can’t remember the name or you’re not sure what you’re in the mood for, tap the Alexa button on the Fire TV remote and ask questions like, “What’s that show about money laundering set in the mountains?” or “Show me British crime dramas with female leads,” and the voice assistant should help you figure it out. It’s the AI equivalent of flipping through the channels, only Alexa is doing the flipping for you.

This is all being powered by a new Amazon-built large language model (LLM) designed to surface movie and TV show content using natural language inputs. It’s starting to roll out to eligible Fire TV devices running Fire OS 6 or higher today. At launch, it’s capable of finding content based on things like topic, genre, plot points, actors, and quotes thanks to being trained on data from services like IMDb.

Amazon’s Joshua Park, senior product manager of Fire TV, demoed AI search to me at Amazon’s Day 1 HQ in Seattle earlier this month. He showed me several queries, including: “Show me the movie where Tom Hanks is a pilot and has to land on the Hudson” (Sully); “What is the TV show that mentions Szechuan sauce from McDonald’s?” (Rick and Morty); and “Show me the nature documentary that was narrated by Obama” (Our Great National Parks). Alexa did a good job with all these — but while it’s neat, it’s all stuff I can Google on my phone while sitting on the couch. 

Amazon does add helpful context to the results, including showing you which apps you have that can stream the show and whether it’s free to you. But what I want from a more intelligent search service is not something to jog my memory but something smart enough to find me something good to watch. I want it to use its vast dataset to sift through the cruft and find me the quality. I want it to be that old-school video store clerk from my youth. 

When Daniel Rausch, VP of Alexa and Fire TV, demoed the search function onstage at Amazon’s fall event last year, that’s literally what he promised, saying using the feature is “like speaking to a great friend who’s also the world’s best video store clerk.”

His demo involved a far more capable Alexa than the one I saw in Seattle. He asked Alexa to “find some action movies for me,” then was able to continue talking to the assistant to winnow it down to movies he wouldn’t have to pay for, those he hadn’t seen yet (or at least weren’t in his Fire TV watch history), ones that were good for his teenagers, and then finally ask it a context cue: “We like video games, which one should we go for?” It suggested Scott Pilgrim. Now that is very useful. 

Park tells me that kind of in-depth, conversational back and forth is planned for future updates. In my time trying out the current capabilities, I wasn’t able to get it to go beyond two queries before it started to fall apart. It also struggled to offer up more than a couple correct answers for broader queries like “Show me Oscar-winning movies from the 1970s.”

“It’s certainly day one for us,” explained Park when I asked about these limits. “We definitely have a view of what we need to do to improve it, so that no matter what the customer asks, we’re able to find the right content for them.”

What it does do well is improve on the current state of Alexa voice search, which — like most voice commands — requires specific nomenclature to surface the right results. With the new Fire TV search, I could chat conversationally with Alexa, including pauses and ums and ers, and it (mostly) understood what I asked. 

But I was largely disappointed in the results. To see if it could help with my family’s viewing situation, I suggested the prompt “Show me some dark comedies with violence.” (I love romantic comedies, and my husband loves horror movies.) It offered up Heathers, American Psycho, Pulp Fiction, and Barbie. Besides Barbie being totally out of left field, the others were all over 20 years old. Not helpful.

Next, I tried something much more specific. We like to find series we can binge-watch together, so I asked, “Show me TV series with more than six episodes that are highly rated.” It suggested two shows, both anime. One was rated a nine out of 10 but the other was a five out of 10. Even for an avid anime fan, that isn’t a great result. 

At this point, I decided to go from what I thought would be a softball question. The kind of thing I might have asked that video store clerk: “Show me something good to watch.” The results were… bizarre. Its first suggestion was Miss Marple (a classic British detective show that I do actually love, but is very old), but its second and third options were The Curious Female and Super Vixens, which not only appear to be ’70s soft-core porn but have very poor ratings on IMDB.

Yes, it’s still early days indeed.

Amazon says the search function is designed to be personalized to you, and the Fire Stick I was using at Amazon’s HQ obviously wasn’t. I just got the new update on my Fire Stick this morning, so was able to repeat the “something good to watch” query and am happy to say there was no sign of curious females. Instead, Alexa suggested Dune: Part Two, Shōgun and Sugar. So, it looks like I might be set for my weekend viewing.