It’s Time for Games to Have Dynamic Dialogue
Last night, I was playing the hot mess that’s the GTA Trilogy, doing the usual things: jacking cars, blowing up cops, chasing that glorious 5 star fame. Just another normal day in Liberty City.
However, things got pretty old by the 6th hooker I chainsawed. You can only listen to the same old “Hey, watch it” gripe, followed by the same hooker Wilhelm scream, before things get boring.
Which got me thinking: why don’t games have dynamic dialogue? And by dynamic, I mean voice lines that change, instead of the same 4 or 5-line crap we get for each NPC.
After all, we’re in the age of voice assistants and adaptive responses. We have Siri on our phones, Alexa and Google at home, and Cortana on our laptops (before Microsoft finally killed it).
More than glorified assistants and smart home keepers, these voice AI are dynamic: they are witty, they are learning all the time, and sometimes the responses they give can be totally and pleasantly surprising.
So why can’t we do the same for games?
Consider the following benefits if games and NPCs had dynamic lines:
Replayability
Some games are renowned for their replayability factor, even after you’ve exhausted the single-player campaign. The classic SWAT 4, for instance, has a randomizer baked in the campaign so that no two missions are the same: the environment can subtly differ, enemies can appear in different spots, and the tactical situation can dramatically change with each playthrough.
Imagine doing that, but with dialogue. No longer do you have to listen to the same “Kirov reporting” and “Bombardiers to your stations” in Red Alert, you can get new responses like “Dropping our load” or “Greetings from Mother Russia”. I’d happily download it again just to hear comradeski Kirov growling new lines while bombing out the Allies.
“Stalin has a present for you.”
Keeping Up to Date
Another pro would be updated dialogue that keeps up with real-world events. Imagine a GTA Online where, instead of the pedestrians saying “Watch it pal”, you hear snark like “Social distancing man, you know it right?”
Then there’s the season. Some games like Fortnite are famous for riding on seasonal trends like Halloween and Christmas. Instead of being constrained to graphical updates, the dialogue should reflect the season. Going back to GTA, it would improve realism if a rival gang member spouted “Merry Christmas asshole” just before punching CJ in the face. A friendly Yule greeting, in keeping with the season.
Virality
What do the most viral Youtube videos of GTA Trilogy and Cyberpunk 2077 have in common?
Answer: they’re all about glitches. More specifically, the bugs people encounter in these half-assed games.
Here we see a formula for viral content: funny, surprising, and totally unexpected. Whether it’s Big Smoke becoming anorexic, or V suddenly assuming the T-pose (sans pants), there’s something hilarious about seeing random surprises in the middle of serious gameplay.
You know what else can deliver that, apart from unintentional glitches? Simple: add more dialogue.
If NPCs could dynamically generate responses like Alexa, you can bet your ass there would be hundreds of Youtube clips chronicling all the responses that Judy Alvarez gives whenever you try to make a pass at her in Cyberpunk.
But of course, dynamism is one thing. There are still roadblocks to turning games to the same level as voice assistants. These include:
Constant cloud connection — Siri and Alexa need to be connected to the cloud to work perfectly. Take away the WiFi and the devices will be greatly handicapped in their responses and functionality. This will also apply to games.
One way to get around this is to offer dynamic dialogue only in online games, such as GTA Online, since they require constant connection.
One upside is that shunting the voice lines off to the cloud will help cut down on install size. The massive Elder Scrolls, for example, takes up 20gb just for voice files, out of its 80gb bulk.
And you can’t even understand most of those voice packs.
Continuous updates — Another obstacle is that generating new responses means ongoing game development. This will be a deal-breaker for many games, where the studio’s involvement ceases upon release, aside from the occasional patch or update.
However, popular games with massive fan bases and remain in continuous development can incorporate dynamic dialogue generation, like No Man’s Sky and Microsoft Flight Simulator.
Fidelity — Like faces, voices have a distinct identity. In GTA V, it’s what sets Michael’s mafioso character apart from the gruff and slightly deranged personality of Trevor.
To make dynamic responses believable, the voice has to be as close as possible to the original. The booming voice of the aircraft carrier captain in RA3 needs to stay the same, as well as the distinctly Aussie twang of the IFV unit.
Short of tracking down the original voice actors and offering them new contracts, this could be solved with technology through voice alteration.
Finally, it’s worth noting that the three biggest voice AI players in the market are themselves branching out into video games: Amazon with their own studio and Prime gaming, Apple with their Arcade, and Google’s Stadia.
These companies have the technology, voice AI expertise, bandwidth, and most importantly, the deep pockets to make it a reality.
So make it happen you guys. And if any of you bite into my idea, you know where to send the check.