The legendary creator behind Ultima discussed his nearly four-decade connection to Dragon Con, the possibility of returning to the Ultima universe, artificial intelligence in game development, the problems facing modern gaming and how traveling to space fundamentally changed his perspective on Earth.

Garriott’s history with Dragon Con stretches back to the convention’s beginning. He recalled being the computer guest of honor at the first Dragon Con and becoming a fan of the convention through that experience.

Those early conventions also put him in the same rooms as creators who had influenced his career. Garriott recalled meeting Dungeons & Dragons co-creator Gary Gygax, whose work he described as fundamental to his own career. He later met Tracy Hickman at Dragon Con 2, beginning a relationship that eventually grew into collaborations that continue today.

For Garriott, conventions offered something else that was difficult for a game designer to find: direct contact with players. Publishing a game wasn’t like performing live onstage, so meeting players became an important feedback loop. At the same time, Dragon Con gave Garriott the opportunity to simply be a fan alongside everyone else.

A Return to the World of Ultima

One of the biggest subjects of the conversation was the future of Ultima.

Garriott directly addressed reports surrounding his efforts to reclaim copyrights associated with his original work. He said he notified the U.S. Copyright Office and Electronic Arts of his intention to reclaim those copyrights, explaining that the 35-year timeline following the 1992 sale reaches 2027.

“Next year I will own all my original copyrights,” Garriott said.

The situation, however, isn’t as simple as saying Garriott will control everything associated with Ultima. He stressed that copyrights and trademarks are separate, with trademark issues still complicating the situation.

Garriott said negotiations are underway concerning what happens next.

But when asked whether that could eventually mean returning to the series, he didn’t hide his intentions.

“We may return to the world of Ultima. That’s the plan.”

What would a modern Lord British game actually look like?

Garriott’s answer suggests it wouldn’t simply be an old Ultima with better graphics.

Ethics and virtues would again play a role in the story, while the world itself would be built around interaction. Garriott described a philosophy in which objects aren’t simply scenery. If players can see something, they should reasonably expect it to work.

Cups and plates could be picked up, placed somewhere else, filled and used. Even mundane objects should behave as though they belong in a functioning world.

“Everything will be interactive,” Garriott said.

That philosophy extends to the world itself.

Garriott isn’t interested in creating enormous environments simply for the sake of scale. Towns should contain enough buildings to make players believe they’re in a town, but those buildings should have a reason to exist.

He similarly favors carefully handcrafted environments rather than depending entirely on procedural generation. A world that looks natural isn’t necessarily a world that produces a good story.

Exploration is another element Garriott believes modern RPGs sometimes sacrifice in the name of convenience.

Many contemporary games give players a destination marker and sometimes a route leading directly toward it. Garriott sees value in making the player discover that route.

He described the example of traveling between two towns separated by mountains and forest. Instead of following a marker directly to a mountain pass, the player might need to enter the forest, reach the mountain range and search along it until discovering the passage.

That discovery is part of the adventure.

When the game shows the player exactly where to walk, Garriott said, “You’re not really exploring to get there.”

That doesn’t mean he wants to preserve every inconvenience from early computer RPGs. Garriott acknowledged that returning to some older keyboard-driven Ultima games can be extremely difficult today and said modernized versions would need modern interfaces.

The goal would be to remove unnecessary friction without removing discovery.

That experimental approach has been part of Ultima from the beginning.

Garriott recalled how the team behind Ultima Online once attempted to build an actual virtual ecology. Grass provided resources for herbivores, while carnivores hunted the herbivores. If players killed too many deer or sheep, predators could move toward populated areas, potentially attacking villagers and creating quests.

In controlled conditions, Garriott said the system worked beautifully.

Then players arrived.

They killed everything.

Wolves, sheep, deer — virtually anything players could attack disappeared faster than the simulation could compensate. Eventually, the development team removed the entire system.

“We spent, you know, human years on it and it was useless,” Garriott recalled.

Other experiments were far more successful.

Dialogue systems and NPC schedules helped create characters who appeared to live independently of the player. Characters could wake up, go to work, eat and eventually return to bed. Others might secretly leave at night and engage in activities players would only discover by abandoning their normal routine and following them.

Those experiments helped create the kind of living virtual world Garriott still wants to pursue.

AI, AAA Development and the Changing Game Industry

Artificial intelligence could play a role in building that next generation of virtual worlds.

Garriott confirmed that his team is already experimenting with AI as a development and prototyping tool. Rather than spending months committing resources to an idea, developers could potentially test concepts quickly before deciding which direction deserves full production.

That could include fundamental questions such as whether combat should be real-time, turn-based or somewhere between the two.

Garriott isn’t approaching AI without concerns.

“As a human, I’m actually profoundly worried about what the future of humanity is going to be like,” he said while discussing the possibility of AI eventually becoming more intelligent than humans.

He also raised concerns surrounding energy and water consumption, training data and whether creators are being compensated for material used to train AI systems.

Within game development, however, he sees the technology as potentially giving individual creators more direct control.

Garriott remembered a time when he could personally create much of a game. As projects grew, his job became increasingly managerial. Instead of making the map, he became the person setting the vision. Leads worked beneath him, with additional teams beneath those leads.

Eventually, the people actually creating something could be several layers removed from the person who originally envisioned it.

“That’s frankly not fun,” Garriott said.

He also believes it can affect the finished product because what ultimately gets created may no longer feel like the work of the original creator.

AI-assisted tools could potentially reduce that distance.

A designer might create a detailed prototype before handing it to an artist or specialist for the final production work. The idea wouldn’t have to pass through as many layers of interpretation, and smaller teams could test ideas far more quickly.

Garriott believes that could make development less expensive while returning creators closer to the act of creation itself.

Cost is particularly important because Garriott believes the modern game industry is facing serious economic problems.

He compared the current environment to the video game crash of the early 1980s, arguing that the traditional AAA model has become increasingly difficult to sustain as development budgets and production timelines have grown.

He also criticized the industry’s increasing dependence on recurring monetization.

Garriott argued that free-to-play games and similar models can encourage developers to build mechanics around repeated purchases and unpredictable rewards. He compared some of those systems to the psychology behind gambling.

His concern comes down to incentives.

When a developer’s revenue comes primarily from selling the game, Garriott argued, the financial goal is to create something worth purchasing. When revenue depends on keeping players spending after they begin playing, developers have an incentive to encourage behaviors that produce additional transactions.

Garriott believes that difference can ultimately influence the design of the game itself.

From Lord British to the Edge of the World

Garriott’s fascination with exploration isn’t limited to virtual worlds.

His father was an astronaut, while his mother was a naturalist and artist. Garriott credits both with helping develop his curiosity at a young age. His mother would take him into the woods to investigate nature, follow insects and understand how things worked.

That desire to explore eventually took him far beyond the woods — to the North Pole, South Pole, the depths of the ocean and orbital space.

Space in particular changed him.

Garriott described orbiting Earth at roughly 17,000 miles per hour, circling the planet approximately every 90 minutes and watching a sunrise or sunset roughly every 45 minutes.

“You cannot take your eyes off the window,” he said.

Looking down at Earth, Garriott began noticing things that are difficult to comprehend from the ground.

The atmosphere appeared remarkably thin. Freshwater seemed limited. He could see deforestation, wildfires and pollution stretching across enormous areas.

Then he passed over Texas.

Garriott recognized his hometown, familiar roads and places where he had grown up. But after spending days observing the entire planet, those familiar locations suddenly existed within a scale he could directly comprehend.

The realization gave him goosebumps.

Earth went from feeling almost impossibly large to what Garriott described as “extremely finite and deeply interconnected.”

Only afterward did he learn that this shift in perception has a name: the overview effect.

Despite everything he has already accomplished, Garriott isn’t finished exploring.

Shortly before Dragon Con, he returned from Indonesia, where he had been diving around Komodo Island. He saw Komodo dragons for the first time and explored a coral reef that had been affected by a recent earthquake.

There are still places on Earth he wants to explore and others he hopes to revisit with his children.

Beyond that, Garriott wants to leave Earth orbit again.

The Moon and Mars are on his list.

Whether the next adventure is on Earth, in space or inside a virtual world, Garriott said he still has a “very long bucket list.”

And that list now appears to include Ultima once again.

Near the end of the conversation, Garriott revealed that one of the first projects being considered is essentially a return to Ultima IV: Quest of the Avatar.

“The first thing we’ll bring back is basically Ultima 4, Quest of the Avatar,” Garriott said, adding that experiments surrounding the project are already underway.

For someone whose career has repeatedly crossed the boundaries between games, technology and exploration, a return to Ultima would bring Garriott back to the world that made Lord British famous.

But if his comments at Dragon Con are any indication, he isn’t interested in simply recreating the past.

He wants to experiment with what comes next.

Photo credit: OpenClaw

Key Takeaways

– AI is moving beyond one-shot prompts and into agent-style tools that can handle multi-step work.
– Big players are racing to turn AI into something that can browse, summarize, automate, and take action.
– That shift could change how people work online, but it also raises bigger questions about trust, accuracy, and control.
– The next AI fight is not just about who has the smartest model. It is about who builds the most useful assistant.

For a while, the AI conversation was all about who had the flashiest chatbot. Bigger model. Faster answers. Better reasoning. More wow-factor demos. That race is still happening, sure, but the real trend now is something a lot more important.

AI is moving from talking to doing.

That is why AI agents are suddenly everywhere. Tech companies are no longer pitching AI as just a tool that answers questions. They want it to handle tasks, chain together actions, search across tools, summarize what matters, and in some cases actually get work done with less hand-holding.

That shift is a big deal because it changes the entire value proposition. A chatbot can impress you for five minutes. An agent that saves you an hour every day is something else entirely.

Why AI agents are suddenly the center of everything

The reason this trend is taking off now is pretty simple. The major AI players have already proven they can build models that generate text, code, and images at a high level. The next step is making those models useful in the messier real world.

That means letting AI interact with software, tabs, files, notes, spreadsheets, research sources, and all the other junk people deal with every day. Instead of asking a model one question at a time, users increasingly want something that can help manage the whole workflow.

That is where the agent framing comes in. AI agents are being sold as systems that can plan across multiple steps, use tools, and complete tasks with less babysitting. In theory, that could make them feel less like fancy autocomplete and more like actual digital workers.

And honestly, that is the pitch that could finally move AI from novelty into habit. People do not stick with tech because it sounds futuristic. They stick with tech because it saves time and reduces friction.

The upside sounds huge, but the risk is real too

There is a reason companies are leaning hard into this. If AI agents work the way their creators want, they could reshape how people search, write, organize projects, shop, code, and communicate.

But that is also where things get tricky fast. The more power an AI system has, the more damage it can do when it gets something wrong. A bad answer in a chat window is annoying. A bad action taken across your files, messages, calendar, or browser is a whole different level of problem.

So this is where trust becomes the real battleground. It is not enough for an AI agent to feel clever. It has to be dependable. It has to know when to stop. And it has to make it obvious what it is doing before it starts pressing buttons on your behalf.

That is probably why the companies pushing hardest into agents are also obsessed with guardrails, approvals, memory, and tool permissions. The moment AI goes from suggesting to acting, the stakes jump immediately.

This is the next real competition in AI

The most interesting part of this whole trend is that it changes what winning looks like. The smartest raw model does not automatically win if a rival product is easier to use, better connected to real tools, or more trustworthy in day-to-day work.

That means the next AI war is not just about benchmarks anymore. It is about product design. It is about ecosystem. It is about whether your assistant can actually fit into someone’s real life without becoming another thing they have to manage.

And that is why AI agents feel like such a major turning point. They push the whole industry toward utility instead of just spectacle.

We are still early, and plenty of these tools are going to overpromise before they settle down. That part feels inevitable. But the direction is obvious now. AI is no longer trying to just sound smart. It is trying to become useful enough that people build daily routines around it.

If that happens, the next era of AI will not be defined by chat alone. It will be defined by who builds the assistant people actually trust to do the work.

Speaking at The Wall Street Journal’s “The Future of Everything” Conference, OpenAI’s Chief Operating Officer, Brad Lightcap, indicated that formal deals with movie studios are not imminent. As reported by TheWrap, Lightcap’s comments, made during an interview with WSJ reporter Keach Hagey, highlight a strategy centered on foundational understanding and collaboration.

“To start any conversation with AI, it starts with an education process. This stuff’s all really new,” Lightcap stated. He further elaborated on the importance of mutual understanding before formal engagements: “There’s a baseline level of trust and understanding that you have to build with partners before you engage just as to what it is that you’re doing.”

This cautious stance comes even as OpenAI has successfully brokered content licensing deals with several major publishers. However, Lightcap explained that the landscape with media companies, particularly in Hollywood, requires a different approach due to the nascent stage of AI in filmmaking and the need to clearly demonstrate its value and address concerns.

Referring to Sora, OpenAI’s text-to-video AI model which has garnered significant attention for its potential in content creation, Lightcap acknowledged its current limitations for professional productions. “People can play with [Sora], but it’s not meant for production-level content,” he remarked. “It’s not, in my opinion, good enough for that yet.”

Despite this, Lightcap expressed optimism about the future trajectory of these tools. “So will it get better? Yeah, of course. But it’s just [about] starting a dialog and starting a process of understanding [with studios currently],” he added, signaling a long-term vision for AI’s role in the entertainment sector.

OpenAI’s current focus, therefore, appears to be less on immediate deal-making and more on fostering a receptive and informed environment within Hollywood, ensuring that future collaborations are built on a solid foundation of trust and shared understanding of AI’s evolving capabilities.