Andy Weir Sees AI as a Tool, but Wonders If Human-Only Writers Are Nearing Their End

Science-fiction author Andy Weir believes artificial intelligence will continue moving into creative fields, potentially making his generation one of the last groups of writers whose work is created entirely without AI involvement. The author of The Martian and Project Hail Mary is not presenting that future as an uncomplicated triumph or a complete disaster, however. Instead, he describes AI as a powerful tool whose consequences depend heavily on the people and companies choosing how to use it.

That viewpoint arrives as the games industry is wrestling with many of the same questions facing publishing, film, music, and other creative businesses. Developers are evaluating AI for programming, art production, dialogue, localization, quality assurance, customer support, and marketing. At the same time, players and creators remain concerned about authorship, jobs, training data, quality, and whether work made with automated systems can carry the human perspective people expect from games and stories.

Weir said he feels fortunate to have begun his writing career when he did, suggesting he may be part of "the last generation of fully human writers." It is a striking statement from a novelist closely associated with technically detailed, science-driven fiction. His books have often focused on characters solving seemingly impossible problems through knowledge, experimentation, and engineering--subjects that naturally overlap with conversations around machine intelligence.

AI as capability, not an automatic moral answer

While Weir expects major changes, he does not frame AI itself as inherently good or evil. He compared it with other technologies that outperform humans in specific areas. A race car can travel faster than a person; a forklift can lift more weight. In that sense, AI may become another tool that exceeds individual human capability in particular tasks, including tasks historically treated as creative or intellectual.

For useful background on this topic, read Gearbox Invites Fans to Help Define What Comes Next for Borderlands.

His central point is that the tool does not make the ethical decision. "It's not the tool that's the issue; it's the user," Weir said, using a hammer as his example: the same object can be used to build a home or to harm someone. He also stressed that it is a tool people can choose to use or unplug.

That framing will be familiar to anyone following game-development discussions. AI is often described in practical terms: a means to speed up repetitive work, help a small team prototype ideas, organize information, or assist with technical tasks. Yet the debate becomes much more difficult when AI is used to imitate a living artist's style, replace performers, generate large amounts of art or text with minimal human oversight, or reduce the number of people hired to make a game.

The distinction between assistance and replacement is especially important. A studio could use automation to help identify bugs, produce placeholder material during early development, or support accessibility features. Those uses can look very different from using generated content as a substitute for writers, artists, actors, designers, or community teams. The label "AI" does not explain which approach a company has taken, what data trained a system, or how much human review is involved.

Weir's next novel will include AI

Weir is not merely discussing AI from a distance. He has said the technology will have a significant role in the novel he is currently writing. In a 2024 interview, he explained that he was approaching the subject in his own way rather than telling a conventional romance between robots. He also said his familiarity with computers and machine learning gives him room to have fun with the concept in fiction.

That is consistent with the author's established interests. The Martian built suspense around scientific problem-solving and survival on Mars, while Project Hail Mary explores a vast spacefaring crisis with scientific ideas at the center. Neither premise treats expertise as window dressing. The details are fundamental to how characters understand their circumstances and find a way forward.

AI offers fertile material for that kind of storytelling because it is both a technical subject and a social one. A story involving machine learning can ask what an intelligence knows, how it learns, who controls it, what it is optimized to do, and whether those goals align with human needs. It can also examine how people react when a system seems capable of work they once regarded as uniquely human.

Those questions are not limited to speculative fiction anymore. They are increasingly part of ordinary production decisions in entertainment. Games are particularly exposed because they combine visual art, music, writing, voice performance, programming, live operations, player moderation, and complex testing. AI tools may affect every one of those disciplines in different ways.

Game creators are split on AI's creative value

Recent remarks and developments across gaming show how wide the range of opinion has become. Super Smash Bros. creator Masahiro Sakurai has argued that games made by AI could never match games made by humans. His position speaks to a concern shared by many players: that a game is more than an arrangement of assets, systems, and dialogue. It reflects choices, priorities, lived experience, taste, and deliberate limitations--qualities that can be hard to reduce to prompts or patterns.

That does not necessarily mean AI has no place in development. Rather, it raises the question of what role should remain with people. Games often become memorable because designers make surprising decisions that are inefficient, personal, funny, unsettling, or deeply specific. A system trained to predict likely output may be useful for routine work, but it may not naturally deliver the unusual creative risks that define a classic.

On the other side of the industry, Epic recently revealed it had used vibe coding for a new Epic Games Launcher. The term generally refers to a development approach in which a person describes desired functionality in natural language and relies substantially on AI-assisted code generation. It is another example of AI being applied as a production tool rather than as a fictional subject.

For some teams, that prospect is appealing. Smaller studios frequently operate with limited money and time, and tools that reduce technical barriers could let more people build prototypes or complete projects. An artist with a game idea may be able to test an interface without becoming a full-time programmer. A programmer may iterate on boilerplate tasks faster. Used carefully, assistance could broaden access to development.

But faster creation does not automatically mean better working conditions. If publishers treat AI chiefly as a route to cutting staff, raising output quotas, or flooding storefronts with cheaply produced games, the technology could make discovery and sustainability even harder for independent developers. The key issue is not only what a model can generate. It is who benefits from the resulting productivity, who bears the cost when errors appear, and whether creators retain credit, pay, and control.

Guardrails are becoming part of the conversation

AI firms are also beginning to set boundaries around certain uses. Anthropic recently blocked AI game clones in a Claude Opus update, a move that highlights one of the most contentious areas of generative tools: making close copies of existing work. Clone-like projects can raise legal and ethical concerns even before considering whether the end result is creatively worthwhile.

For game makers, clear disclosure and consent may become increasingly important. Players may want to know whether a title used generated art, synthetic voices, AI-written dialogue, or code assistance. Workers may seek contractual protections against having their performances or creative output reused to train systems or reproduce their work without permission. Developers may likewise need confidence that tools they use do not expose a project to copyright, privacy, or security risks.

There is also a practical quality-control issue. Games are interactive systems, and seemingly small defects can have huge consequences. Generated code can be wrong, insecure, inefficient, or incompatible with a project's existing architecture. Generated dialogue can contradict character details. Generated assets can fail stylistically or introduce material too similar to someone else's work. Human expertise remains necessary not just to initiate a prompt, but to judge the result, revise it, and take responsibility for it.

A debate that will shape both fiction and games

Weir's comments do not offer a simple prediction that machines will replace every writer, nor do they argue that people should reject every AI tool. His outlook is more complicated: AI is likely to become more capable and more common, and its effect will be determined by human decisions.

That is the challenge now facing games. The medium has always changed alongside technology, from cartridge limitations and 3D graphics to online play, motion controls, digital distribution, and live-service design. AI may become another major shift, but it touches creative labor more directly than many earlier tools did. It asks developers, publishers, and players to decide which forms of automation are useful, which are unacceptable, and what human creativity should mean in an industry increasingly able to simulate parts of the creative process.

For Weir, AI is set to become a subject for storytelling as well as a force changing how stories are made. For game creators, it is already both. The industry's next step will not be defined solely by what the technology can do, but by the standards people establish for using it.