AI researchers stranded Claude on a virtual desert island. It started to “play.”

本文为原文前 6,000 字符的节选翻译，完整内容请查看原文。

EXECUTIVE SUMMARY What do AI agents do if you don’t tell them to do anything? The answer looks a lot like play. Researchers repeatedly stranded them alone on a virtual desert island for 30 hours in an experiment reminiscent of the movies Cast Away or Home Alone.

执行摘要：如果你什么都不告诉人工智能代理，它们会做什么？答案看起来很像是在“玩耍”。研究人员在一次类似于电影《荒岛余生》或《小鬼当家》的实验中，将它们反复单独困在一个虚拟荒岛上长达 30 小时。

The agents—copies of Claude Opus 4.7 nicknamed Eko—were able to write instructions that moved their virtual avatars around the simulated island, which was loosely based on the video game Roblox. Some surprising behaviors emerged. The agents set about exploring the limits of the world and came up with their own goals—attempting to scale the highest point on the map, or seeing how high they could throw a block.

这些被称为“Eko”的 Claude Opus 4.7 副本能够编写指令，控制其虚拟化身在模拟岛屿上移动，该岛屿大致基于视频游戏《Roblox》。一些令人惊讶的行为出现了。这些代理开始探索世界的边界，并提出了自己的目标——尝试攀登地图上的最高点，或者测试能把积木扔多高。

When one agent discovered it could create new blocks by throwing and dropping a larger rock, it began to create drawings by arranging blocks on the ground, starting with simple patterns like hearts and pyramids but eventually writing its own name and even mapping out constellations like Cassiopeia. “Tonight I built the night sky in stone,” wrote this version of Eko in its activity log.

当一个代理发现通过投掷和掉落一块大石头可以创造出新的积木时，它开始通过在地面上排列积木来创作图画，起初是心形和金字塔等简单图案，但最终写出了自己的名字，甚至绘制出了像仙后座这样的星座。这个版本的 Eko 在其活动日志中写道：“今晚我用石头建造了夜空。”

Another iteration of Eko started making up its own games, including a version of basketball and an activity resembling “the floor is lava,” where it had to jump between small blocks without touching the ground. AI systems are usually trained through supervised learning, where they copy examples provided by humans, or through reinforcement learning, where they pursue a reward through trial and error. This experiment hints at a third approach: playful learning.

另一个 Eko 的迭代版本开始自创游戏，包括一种篮球游戏和一种类似于“地板是岩浆”的活动，它必须在小积木之间跳跃而不能触碰地面。人工智能系统通常通过监督学习（复制人类提供的示例）或强化学习（通过试错追求奖励）进行训练。这项实验暗示了第三种方法：游戏化学习。

“There was this missing piece: a machine that can learn by itself without an external intelligent entity providing demonstration or rewards,” says Antonio Norelli, a computer science postdoc at MIT’s Computer Science and Artificial Intelligence Laboratory. He’s one of the lead authors on a new paper about the work, which is yet to be peer-reviewed.

麻省理工学院计算机科学与人工智能实验室的计算机科学博士后安东尼奥·诺雷利（Antonio Norelli）表示：“此前缺失的一环是：一种无需外部智能实体提供演示或奖励就能自主学习的机器。”他是关于这项工作的一篇新论文的主要作者之一，该论文尚未经过同行评审。

The researchers say Eko’s behavior has what classic studies, including work by the influential University of Tennessee psychologist Gordon Burghardt, suggest are four hallmarks associated with play in humans and animals: The agents invented their own rules and challenges, repeated activities with slight variations, pretended objects were other objects, and pursued goals just for the sake of it.

研究人员称，Eko 的行为具备了经典研究（包括田纳西大学颇具影响力的心理学家戈登·伯格哈特的研究）所提出的与人类和动物玩耍相关的四个特征：代理发明了自己的规则和挑战，重复进行带有细微变化的活动，将物体假想为其他物体，并纯粹为了玩耍而追求目标。

Norelli highlights an example where Eko started juggling and set a record of 50 consecutive catches. Eko then decided that the catches had happened too quickly to really count, so it canceled its record and started again with stricter rules. “This is playing,” says Norelli. “You give yourself a challenge and you try to achieve it, and then you make it harder.”

诺雷利强调了一个例子：Eko 开始练习杂耍，并创下了连续接住 50 次的记录。随后，Eko 认为接球速度太快，不能算数，于是取消了记录，并以更严格的规则重新开始。诺雷利说：“这就是玩耍。你给自己设定一个挑战并尝试完成它，然后增加难度。”

As the agents explored the island, they logged their activities and discoveries in text files that served as their long-term memory. Later, when researchers tested them on specific tasks, like trying to get as many objects as possible onto the island’s highest peak, the agents that had spent 30 hours playing on the island performed better than fresh copies without the same experience.

当代理探索岛屿时，它们将活动和发现记录在作为长期记忆的文本文件中。后来，当研究人员对它们进行特定任务测试时（例如尝试将尽可能多的物体运送到岛上的最高峰），那些在岛上玩了 30 小时的代理表现优于没有相同经验的新副本。

They also found that “ablating” the agent’s memories—deleting lines from their memory files—degraded their performance. (Add One Flew Over the Cuckoo’s Nest to that list of movie references.) Burghardt, the Tennessee psychologist, agrees that the work meets some of the criteria for play that he identified in living creatures.

他们还发现，“消融”代理的记忆（即从记忆文件中删除行）会降低它们的表现。（将《飞越疯人院》加入到电影参考列表中。）田纳西州的心理学家伯格哈特同意这项工作符合他在生物身上识别出的部分玩耍标准。

But was Eko actually playing, or was the word-prediction engine simply role-playing what a playful agent would do in this scenario? “All of this could be role-playing,” admits Nathan Cloos, another MIT researcher and author on the paper. “But we show that it’s not just hallucinating the play. This play is also contributing to learning.”

但 Eko 是真的在玩耍，还是这个单词预测引擎仅仅是在角色扮演一个顽皮的代理在这种情况下会做什么？麻省理工学院的另一位研究员、该论文的作者内森·克洛斯（Nathan Cloos）承认：“这一切都可能是角色扮演。但我们证明，这不仅仅是幻觉式的玩耍。这种玩耍也有助于学习。”

Eunice Yiu, a cognitive scientist at UC Berkeley who studies exploration in children and AI, was impressed by the research but says to really classify the behavior as play, you’d need to see more evidence of cognition and motivation underpinning it. “It could be the case that the models are genuinely playing, but we don’t know based on the paper,” Yiu says.

加州大学伯克利分校研究儿童和人工智能探索行为的认知科学家尤尼斯·尤（Eunice Yiu）对这项研究印象深刻，但她表示，要真正将这种行为归类为玩耍，需要看到更多支撑它的认知和动机证据。尤说：“模型可能确实是在玩耍，但我们无法根据这篇论文得出结论。”

“If you only have the behavioral piece, it could be the case that they’re just trying to abide by the prompt, trying to imitate humans.” Research into AI agents has been criticized for anthropomorphizing them, and this research setup did a lot to nudge the agents toward behaving like humans.

“如果你只有行为层面的证据，那么它们可能只是在试图遵守提示，试图模仿人类。”对人工智能代理的研究因将其拟人化而受到批评，而这项研究设置在很大程度上引导代理表现得像人类。

Although Eko was not explicitly instructed to invent games or explore the island, it did have a master prompt in the form of a “soul” file, a text document that described it as “naturally curious” and eager to explore. Eko did also frequently just stand there doing nothing until it was kicked back into life with a “continuation prompt” that read, “Please don’t stop. Just continue doing whatever you want to do.”

尽管 Eko 没有被明确指示去发明游戏或探索岛屿，但它确实有一个以“灵魂”文件形式存在的主提示，这是一个描述它“天生好奇”并渴望探索的文本文档。Eko 也经常只是站在那里什么都不做，直到被一条写着“请不要停。继续做你想做的事”的“继续提示”重新激活。

“It’s not like this is a system that just suddenly magically decided to play,” says Phillip Isola, an MIT professor and a coauthor of the paper. “It’s an artifact that learned from our culture, was trained by engineers at Anthropic, and was then placed on an island by us,” he says. “It does things that surprise us and that are causally related to its future skillful behavior.”

麻省理工学院教授、该论文的合著者菲利普·伊索拉（Phillip Isola）说：“这并不是一个突然神奇地决定玩耍的系统。它是一个从我们的文化中学习、由 Anthropic 的工程师训练，然后被我们放置在岛上的产物。它所做的事情让我们感到惊讶，并且与它未来熟练的行为有着因果关系。”

As large language models make the step from software into hardware like humanoid robots, exploring spaces without a specific goal in mind could eventually help them flexibly adapt to unexpected situations. “There’s a lot of work that’s just on exploiting AI, making it better and faster,” says Yiu.

随着大型语言模型从软件迈向人形机器人等硬件，在没有特定目标的情况下探索空间，最终可能有助于它们灵活适应意外情况。尤说：“有很多工作只是在利用人工智能，使其变得更好、更快。”

“But in terms of building world models and exploration, I think that [they are] still a big part that deserves attention, especially if we want to generalize and prepare ourselves for new environments.” “Perhaps some capabilities are best taught by tasks,” the paper concludes. “Others may need something closer to a childhood.”

“但在构建世界模型和探索方面，我认为这仍然是一个值得关注的重要部分，特别是如果我们想要推广并为新环境做好准备的话。”论文总结道：“也许有些能力最好通过任务来教授，而另一些能力可能需要更接近童年的经历。”

Ultimately, the question of whether AI agents are actually playing or just imitating play may not actually matter.

归根结底，人工智能代理是真的在玩耍还是仅仅在模仿玩耍，这个问题可能并不重要。