I’m the type of person who revels in the second screen experience. Any time I’m watching a TV show, I’m also on my laptop or have my phone in hand. Either before, during, or after, I’ll relentlessly search the web for details about said show. Who played that character? Where else have I seen that actor? Where was The White Lotus season 2 filmed? What did that moment in the finale mean?
Google is my default source for this, as is the case with virtually any query one might have on the internet. But ChatGPT has been making waves recently, so I was curious to see how it works and if it was any easier to use or yielded more palatable, easily digestible answers versus pages and pages of search engine results.
What is ChatGPT?
ChatGPT was created by OpenAI as a chatbot that can answer queries in a conversational form (it can also compose long-form content, including everything from articles to movie scripts, but that’s another angle altogether). Unlike Google, which presents a series of web pages containing information pertaining to the question you have asked or term you have searched, ChatGPT is more like talking with a friend. Ask a question, using any language format you prefer, and it will reply in kind. Like any other conversational AI, ChatGPT is trained to understand conversational language versus keywords.
It uses something called large language models to predict the next word in a series of words you type. This is combined with Reinforcement Learning with Human Feedback (RLHF) to allow it to follow directions and deliver responses that take into account context and exactly what you likely want to know.
You can try ChatGPT by signing up online with an existing or new OpenAI account.
How does ChatGPT works in the context of TV?
I was curious to see how well ChatGPT would work at answering the types of questions I would typically “Google” when watching a TV show or movie. I tried various searches, including the same search one week and then the next, which I found sometimes yielded different results. ChatGPT is not only continuously learning, but also refining.
Localized questions
I started with a localized question: “Where can I watch House of the Dragon in Canada?” ChatGPT initially did not know this when I checked last week, advising that the show is available through HBO Max, which is not currently available in Canada. it suggested checking the HBO Max website or social media to get updates as to when it might be available north of the border.
Interestingly, when I asked the same question a week later, it cited Crave with the HBO add-on, along with rent/purchase options through Google Play, Apple TV, and Amazon Prime Video. The platform is clearly getting smarter over time.
I plugged the exact same query into Google just to do a comparison and Crave popped up instantly as the top result. It advised that House of the Dragon is available exclusively through the streaming service in Canada and quick results provided the premiere date and time along with the new episode release schedule. Score one for Google.
Information about actors
I switched to another common query of mine about casts: I asked, “Who plays David in Schitt’s Creek?” Initially, the answer not only provided the name of the actor (Daniel Levy), but also details about how he co-created the show, when and where it premiered, and even about the show’s multiple Emmy wins. Some of the information, however, was dated. ChatGPT, for example, neglected to inform that Schitt’s Creek was previously available through Netflix and now is on Hulu. When I followed up with “where can I stream Schitt’s Creek?,” however, the current information appeared.
In the span of a week, ChatGPT was already refining and updating information as it learns. It even eliminated dated and extraneous information. The same search the following week simply noted Levy’s name and eliminated all the extra information about the show.
In Google, the results show Levy’s name along with a photo of him and three other actors people also search for (Eugene Levy, Sarah Levy, and Annie Murphy.) With one click, I could get to more search results for Levy, including his Wikipedia and IMDb pages. For someone like me who loves to go down the rabbit hole of searches, this was a far better experience.
Not yet ready for current events
Testing current events, I asked “how many Golden Globes did HBO win this year?” The reply indicated that ChatGPT’s knowledge cutoff is 2021, but it initially informed about HBO’s historically strong presence at the Golden Globes and recommended checking the official Golden Globes website or news outlets that covered the event. No links were provided, and the same search a week later resulted in a much more refined and succinct reply, as per above.
Google provided the straight answer: four wins out od 14 nominations, along with a list of relevant news stories covering the event.
More context needed
I decided to try something vague and asked, “Who is the blonde woman on The White Lotus?” ChatGPT provided details about the show in its response, but recommended I check the cast list on IMDb or other sites to find out because “without further information on the specific episode you’re referring to, it’s hard to tell you exactly who the blonde woman is.” When I tried the same question a week later, the response was nil, advising that further details were necessary to generate an answer: it didn’t even recognize that The White Lotus is a show.
Google, on the other hand, instantly identified Sydney Sweeney, who indeed is a main character on season 1 with blonde hair. When I added “season 2” to the inquiry, it pulled up Jennifer Coolidge, who played Tanya McQuoid-Hunt on both seasons, as the result. This may be because the movie Legally Blonde is listed among her credits and Google picked up the keyword “blonde” alongside her name and credit on the show. The most accurate answer, of course, should have been Daphne Babcock, played by Meghann Fahy.
When I delved deeper and asked “who died on The White Lotus?” ChatGPT didn’t know the answer (again, the database cutoff is 2021). I also had to clarify context, adding “the HBO show” before the title to clarify what I meant by “The White Lotus.” Google, not surprisingly, was happy to deliver a spoiler with the name of the fated character who met their end on season 2.
Further showing ChatGPT’s limitations, I asked for insight on the series The Last of Us and if it is worth watching? The result initially indicated there is no such series based on the video game (one just premiered on HBO on January 15, 2023). The second time I checked several hours later, the engine did recognize the show and that it has gotten generally favorable reviews, but suggested I make my own judgment call as to whether I should watch or not.
Plugging the same question into Google (“How is the first season of the show The Last of Us? Is it worth watching?) yielded several review links from top sites, as well as sources like Reddit.
Recommendations
I am often searching for an interesting new show to watch, either looking online (at sources like Digital Trends!) or talking to family members and friends to see what they are watching. I tried this with ChatGPT to see how well it did with recommendations. I noted that I was looking for a unique comedy, and it provided some worthwhile suggestions. When I dialed down to TV shows versus movies, it provided more ideas, along with a short synopsis of each show and why I might like it. In this respect, ChatGPT is sort of like talking to a friend. But while you’ll get detailed recommendations, you won’t (yet) get current ones.
Will ChatGPT replace Google?
It’s far too early to determine if ChatGPT has the potential to be the new Google when it comes to those second screen searches for TV shows and movies. Currently, there’s still a lot of work to be done. While it’s nice to receive replies in a conversational fashion, it still feels like having a conversation with a robot.
If all you want is a simple answer to a single question, ChatGPT will give you that and (sometimes, though not always) additional facts you may or may not want to know as well. It’s like talking to that friend who has a wealth of trivia-like knowledge about movies and TV shows — always with the answer at the tip of his tongue, as well as fun facts you only sometimes want to hear.
In other ways, the replies effectively send you to Google for the answer because ChatGPT doesn’t have it. There, you’ll find a succinct answer (in most cases) with a line list of additional articles to peruse if, and only if, you want more. The results are instantaneous and succinct, with tons of additional information to look into further if desired.
If you love going down the internet rabbit hole when you search (as I do), you might prefer this type of continuous experience, which you won’t get with ChatGPT without engaging the conversation (for example, asking “what else has Dan Levy been in?” and “Tell me about Dan Levy’s life”) .There aren’t links to click, though you can continue a conversation as long as you like. The conversational aspect is nice, but it takes more time. There’s a flashing square that can be likened to the three dots on a phone messenger app when someone is typing a reply. Once ready to reply, words are typed from ChatGPT in real time, which may not appeal to those looking for instant gratification.
Nonetheless, over time, and as its database of information grows, ChatGPT has potential. With a larger database of information that is continually being updated, more timely and detailed results could be possible. The conversational format, especially when using voice commands, could be more appealing with some more time to adjust to the change.
Bottom line: When it comes to movies and TV and related searches and inquiries, ChatGPT has serious potential for sourcing details and helping those who love the second screen experience like me. But for now, I’ll stick with Google.
