Crazy Taxi: World Tour preview – why SEGA’s AI backlash misses the point

Crazy Taxi: World Tour should have been an easy win. SEGA brings back one of its loudest, most recognisable arcade brands, puts Axel back behind the wheel, sends the series around the world, adds modern multiplayer, keeps an Arcade Mode for purists and promises a new story-led campaign with high-speed missions across five cities. For anyone who remembers the Dreamcast era, that is a tidy pitch. For younger players, it is a rare racing game that does not look obsessed with licence plates, tyre compounds or photorealistic puddles. It wants to be fast, silly and immediate. Instead, the conversation has been hijacked by two letters: AI. The criticism is understandable, at least on the surface. Generative AI has become a toxic phrase in creative industries because it is often associated with cheap production, copied style, job cuts and soulless output. Many players hear that a studio used AI and instantly imagine a boardroom replacing artists with prompts. That fear is not invented. The wider technology sector has done plenty to earn suspicion.

Why the AI criticism feels too absolute

The strongest argument against SEGA is that any use of generative AI normalises a technology built on questionable training data and environmental cost. That is a serious concern, and no publisher should be allowed to wave it away with vague language. Players are right to ask who made the work, whether human artists were paid, whether performers were protected and whether the final game contains generated material.

But the loudest criticism around Crazy Taxi: World Tour often skips those questions and jumps straight to a verdict. It assumes that AI involvement automatically means lazy development. That is not a fair reading of what has been described so far.

That distinction matters. A bad use of AI would be pushing generic machine-made posters, textures, character designs or voice material directly into the finished game and charging full price for it. A more limited use would be asking a tool to produce rough visual prompts, then having artists interpret, reject, redraw and rebuild the actual assets. The first deserves heavy criticism. The second deserves scrutiny, but not automatic cancellation.

Transparency should not be punished

There is another awkward point in this debate: disclosure is being punished. If a publisher says nothing, players may never know what internal tools were used. If a publisher admits limited AI use, the game risks being branded as contaminated before anyone plays it. That creates exactly the wrong incentive.

The games industry needs clearer disclosure, not less. Players should know whether AI helped generate final assets, placeholder references, code suggestions, localisation drafts, marketing copy or internal prototypes. These are not the same thing, and they should not be discussed as if they are.

Crazy Taxi: World Tour has become a useful test case because the reaction shows how immature the public vocabulary around AI still is. Many critics are using one label for several different practices. A game that uses AI-generated voice lines, a game that trains on living artists’ work, a game that uses machine-assisted upscaling and a game that uses early reference images are not automatically equivalent. The ethics differ. The labour impact differs. The creative risk differs. The consumer impact differs.

That does not mean SEGA deserves a free pass. It means the criticism should be precise. Ask for clarity. Ask for credits. Ask for proof that human artists remain central. Ask whether generated work appears in the final build. But do not pretend a reference tool is the same thing as replacing an art department.

The real question is whether Crazy Taxi still feels handmade

What matters most for the player is not whether a tool existed somewhere in the pipeline. It is whether the finished game has personality, intention and craft. Crazy Taxi is not a franchise that can survive on generic content. Its whole appeal is attitude: bold city layouts, ridiculous traffic patterns, aggressive music, jump-friendly streets, exaggerated passengers and that constant pressure to squeeze one more fare into the timer.

If World Tour feels bland, overproduced or algorithmic, the criticism will write itself. If it feels sharp, readable and designed around player expression, the AI debate will become part of the production story rather than the whole story.

From what has been shown and described so far, the game at least understands what it has to prove. The move to a globetrotting structure gives the sequel a clear hook beyond nostalgia. Five cities allow SEGA to vary terrain, architecture and mission rhythm without abandoning the arcade DNA. The campaign premise, with Axel chasing a mysterious organisation after his stolen taxi, is simple enough to stay out of the way while giving the game a reason to move from place to place.

That is exactly the level of story Crazy Taxi needs. Too little narrative and it risks feeling like a remaster with extra maps. Too much and it could drown the pick-up-and-play energy that made the original work. World Tour appears to be aiming for the middle: a framework for spectacle, not a cinematic drama with taxi sections.

Gameplay preview: speed, chaos and the need for precision

The best sign is that SEGA is not presenting Crazy Taxi: World Tour as a realistic racer. It is still about outrageous drifts, big air, messy shortcuts and getting paid for making risky driving look stylish. That is the correct foundation. Crazy Taxi has always lived somewhere between racing game, score attack and urban obstacle course. It is less about clean driving and more about controlled panic.

The challenge will be physics. Modern arcade racers often confuse accessibility with floatiness. Crazy Taxi cannot afford that. The car must be loose enough to perform absurd stunts, but precise enough that expert players can shave seconds off a route. The original games were easy to understand and surprisingly demanding to master. World Tour needs that same friction: instant fun for newcomers, but enough depth for players who want to learn routes, boost timing and stunt chains.

The campaign missions and side jobs could help here. If designed well, they can teach players advanced driving without turning the game into a tutorial. Deliveries, chase sequences, stunt challenges and odd jobs can all push different skills while keeping the pace high. The danger is checklist design. Crazy Taxi should not become a map full of icons. Its missions need to feel like arcade challenges, not open-world chores.

Arcade Mode is the smartest promise

The inclusion of Arcade Mode may be the most important feature. It tells long-time fans that SEGA knows what cannot be lost. A modern campaign is welcome, multiplayer is commercially necessary and customisation gives players a reason to keep playing, but the soul of Crazy Taxi is the timer. Pick up passenger, find route, drive badly in the most profitable way possible, repeat until the clock wins.

Arcade Mode also gives World Tour a clean identity in a crowded market. Many racing games now chase live-service progression, ranked events, garages and seasonal economies. Crazy Taxi has the chance to be refreshingly direct. If Arcade Mode is fast to load, score-focused and built for replay, it could become the feature that keeps the game alive after the campaign credits.

For a computer magazine audience, that matters because the PC version will likely live or die on replayability, performance and input flexibility. A good Arcade Mode is not just a nostalgic extra. It is the part most likely to survive on leaderboards, streams and short-session play.

Multiplayer could be brilliant or messy

Cross-platform multiplayer is the riskiest addition. On paper, it makes perfect sense. Crazy Taxi’s chaos seems made for competitive and co-operative modes, with players fighting over fares, racing to objectives, blocking routes and showing off reckless driving. The concept sells itself.

In practice, multiplayer will need careful structure. If every player is simply thrown into a city at once, the result could become noisy rather than fun. The best version would build modes around readable goals: fare races, stunt contests, team deliveries, route battles and limited-time city events. The worst version would be an unfocused online sandbox where everyone crashes into each other for ten minutes and leaves.

Still, the potential is real. Crazy Taxi has always been social even as a single-player game. It was built for spectators. Friends could watch, laugh, criticise routes and demand one more attempt. Online multiplayer is a natural extension of that energy if SEGA keeps the rules simple and the action legible.

Visual style: bright beats realistic

World Tour’s arcade-style presentation is the right call. Crazy Taxi should not look like a simulation. It needs readable roads, loud landmarks, exaggerated characters and cities that invite dangerous shortcuts. The series is at its best when environments feel like playgrounds rather than believable traffic systems.

This is another reason the AI panic feels premature. The visible art direction matters more than the existence of early reference tools. If the final game has coherent city design, memorable locations and a consistent visual language, then human art direction has done its job. If it looks generic, the criticism will be justified on artistic grounds, not simply because a tool was used somewhere in development.

The soundtrack will also carry heavy expectations. The franchise’s identity is tied to punk energy and late-90s arcade noise. Leaning into that heritage is sensible, but World Tour cannot rely only on recognition. It needs music that supports long sessions without feeling like a museum piece.

Early verdict: promising, controversial and worth judging properly

Crazy Taxi: World Tour is not a finished game yet, so a final scored review would be dishonest. Based on its current pitch, though, this is one of SEGA’s more promising revivals. The concept is clear, the feature list makes sense and the game appears to respect the arcade structure that made the series memorable. The campaign gives it modern shape, Arcade Mode protects the original loop, customisation adds retention and multiplayer could broaden the audience.

The AI controversy should not be ignored, but it should be argued properly. If SEGA ships AI-made content in place of human craft, it deserves criticism. If it used generative tools only as temporary references while human developers created the final game, then the backlash is disproportionate. Players can demand transparency without treating every production tool as proof of creative failure.

For now, Crazy Taxi: World Tour deserves cautious optimism. It has to prove its handling, its mission design, its online structure and its artistic personality. But the idea that limited AI reference use automatically ruins the game is too easy, too neat and not very useful. The better standard is simpler: judge the finished work, credit the humans behind it and keep pressure on publishers to be honest about how games are made.

If SEGA delivers a fast, funny and sharply designed arcade racer, Crazy Taxi: World Tour could be more than a nostalgic comeback. It could be a reminder that old arcade ideas still have room to evolve, even in an industry arguing loudly about the tools used to build them.