Employers are boosting productivity expectations even as workers struggle to translate AI confidence into results, according to a survey of 2,000 workers published Tuesday by WalkMe.根据 WalkMe 周二发布的一项针对 2000 名员工的调查显示，尽管员工们在将 AI 信心转化为实际成果方面面临困难，但雇主们仍在不断提高对生产力的期望。
More than half of respondents said AI has led their manager or team to expect more output in the same amount of time, the adoption platform provider found.这家采用平台提供商发现，超过半数的受访者表示，AI 导致他们的经理或团队期望他们在相同的时间内完成更多的工作产出。
From the top of the executive chain, business leaders are passing down the pressure to perform using AI. More than half of employees said senior leaders don’t fully understand their company’s public AI strategy.从高管层开始，商业领袖们正在向下传导利用 AI 提升绩效的压力。超过半数的员工表示，高层领导并不完全理解公司公开的 AI 战略。
“The higher you go in a company, the more they believe in AI,” WalkMe Global Field CTO KJ Kusch told Channel Dive. “But when you get to boots-on-the-ground, there’s a struggle with how to use it. There’s people who believe in it for context or writing, but I don’t think the real meat of ‘does AI do the tasks that matter the most?’ is really there yet.”“在公司里职位越高，对 AI 的信念就越坚定，”WalkMe 全球现场首席技术官 KJ Kusch 对 Channel Dive 表示，“但当你深入一线时，大家都在为如何使用它而苦恼。有人相信它在提供背景信息或写作方面的作用，但我认为 AI 是否真的完成了那些最关键的任务，目前还很难说。”
Some employees are developing their own strategies to navigate the AI productivity squeeze. One-third of respondents reported pretending to be more skilled in AI than they really are. Another third said they have passed off AI-generated work as their own.一些员工正在制定自己的策略来应对 AI 带来的生产力挤压。三分之一的受访者承认自己假装比实际更精通 AI。另有三分之一的人表示，他们曾将 AI 生成的工作成果冒充为自己的作品。
Yet confidence in AI remains high. Nine in 10 respondents reported feeling confident with AI tools, though only a quarter said AI worked on their first try.尽管如此，人们对 AI 的信心依然很高。十分之九的受访者表示对 AI 工具充满信心，但只有四分之一的人表示 AI 能一次性达到预期效果。
The survey points to a “confidence trap,” Kusch said. Employees say they’re AI-fluent but the results don’t match the claim.Kusch 指出，这项调查揭示了一个“信心陷阱”。员工们声称自己精通 AI，但实际成果却与此不符。
“I’m not sure AI is giving results that really provide value to a company, like driving revenue or finishing business-critical work,” she said. “But employees are using it. They have better presentations than I’ve ever seen, but I’m not sure that actually matters. They write the best performance reviews, they send the best emails, but I think people wish it did more.”“我不确定 AI 是否真的为公司带来了实质性价值，比如推动营收增长或完成关键业务，”她说，“但员工们确实在用它。他们做出的演示文稿比我见过的任何人都漂亮，但我怀疑这是否真的重要。他们写出了最棒的绩效评估，发出了最完美的邮件，但我认为人们其实希望它能做得更多。”
The survey also found that there’s a generational gap in AI confidence. More than 9 in 10 Gen Z workers reported feeling confident in AI, though they were also the most likely to overstate their skills. Nearly half of the Gen Z cohort surveyed admitted they overstated their AI proficiency, compared to 13% of Baby Boomers. 调查还发现，在 AI 信心方面存在代际差异。超过 90% 的 Z 世代员工表示对 AI 充满信心，但他们也是最容易夸大自己技能的群体。近一半受访的 Z 世代员工承认自己夸大了 AI 水平，而婴儿潮一代中这一比例仅为 13%。
“It’s not so much pretending as it is ‘you don’t know what you don’t know,’” Kusch said. “If an AI agent suggests something, you might now know if it’s wrong or right, so you just believe it. I’m not shocked that they’re more confident, because they’re not the ones typically finding errors. It’s the people who have been there, with more work experience.”“与其说是伪装，不如说是‘你不知道自己不知道什么’，”Kusch 说，“如果 AI 代理给出了建议，你可能根本不知道它是对是错，所以就盲目相信了。他们更自信并不让我感到吃惊，因为他们通常不是负责发现错误的人。那些拥有更多工作经验的人，才更容易发现问题。”
Kusch added that one of the most effective strategies for AI implementation she’s seen is using AI characters to act as on-screen, interactable coaches that help employees make decisions rather than instill fear about job security.Kusch 补充说，她所见过的最有效的 AI 实施策略之一是使用 AI 角色作为屏幕上的互动教练，帮助员工做出决策，而不是让他们对工作保障产生恐惧。
“It’s like bringing Microsoft Clippy back,” she said. “Instead of asking employees to put in answers, it asks them to help validate outputs, so the human feels like they’re in the loop of AI. Whereas I think the way AI is mostly delivered is through answers prompts, a bunch of work and an agent telling you whether you’re right or wrong at the end. It’s not a tool; it’s like filling in a ginormous comment box.”“这就像让微软的 Clippy 回归，”她说，“它不是要求员工输入答案，而是请求员工帮助验证输出结果，这样人类就会觉得自己参与到了 AI 的循环中。而我认为目前大多数 AI 的交付方式是基于提示词的问答，一堆工作之后，最后由代理告诉你对错。这根本不是工具，就像是在填一个巨大的评论框。”
