微软年度《未来工作》报告，提出了从远距工作转向人工智能 (AI)，为人工智能的变革力量提供了令人信服的案例，同时也提出了人工智能对劳动力冲击的关键问题。
迹象表明，使用ChatGPT的知识工作者速度提高了37%，工作品质提高了40%，但准确性下降了19%。这表明人工智能可以提高生产力和质量，但可能会牺牲精准度。然而，报告乐观地认为使用者体验（UX）解决方案可以缓解这个问题。
对微软Copilot 365企业用户的调查，进一步确认了人工智能的好处。调查结果显示，73%的用户同意Copilot（AI 办公助手）提高了他们的速度，85%的用户表示它有助于加快文件起草速度，72%的用户表示在一般或重复性任务上花费的脑力更少。这些结果凸显了人工智能优化工作流程的潜力。
有趣的是，人工智能，特别是语言学习模型 (LLM)，对新员工或低技能工人的好处最大，提高了他们工作能力的43%，而高技能员工则只提高了17%。这意味著人工智能可以成为提高劳动力技能和赋权的有效工具，为职业生涯早期的个人提供支持，同时也可是增强经验丰富的专业人员的工具包。
我们需要基于语言学习模型（人工智能的一种）的工具来质疑我们的信念，帮助我们评估情况并提供不同的观点。这意味著，虽然人工智能可以增强我们的能力，但它不应该取代我们批判性思考和评估事物的能力。这个想法是让人工智能像苏格拉底一样，不断鼓励员工更深入地思考并从不同的角度看待事物。
批判性思考之重要
报告也提出了从内容创建到内容分析和整合的重大转变。未来的工作可能需要更多地关注于理解和解释人工智能生成的信息，而不是创建信息。这凸显了批判性思考、领导力和情绪智商等技能日益重要，这些技能是人工智能无法复制的。
人工智能可以将简单的命令分解为微时刻和微任务去执行，从而提高整体品质和效率。报告建议，分析和整合人工智能产生的信息可能比搜寻和创建信息变得更加重要，进而可能会改变人们的工作性质，从内容生产转向分析和整合。
令人关注的是，“替代与增强”的二元论述受到了挑战，反之提出“创新与自动化”新论述作为更细致的框架。报告认为，虽然人工智慧可以自动化某些任务，但其也可以催生全新的角色和产业。关键是监控和促进这种以人为本的创新以及自动化。
最后，报告提出了人工智能增强未来的令人信服愿景。然而，报告也强调，这需要批判性评估、谨慎的执行，以及仔细考量人工智能对职场伦理带来的冲击。
陈奕强《未来工作》原文：Future of Work
Microsoft’s annual Future of Work report pivots from remote work to focus on artificial intelligence (AI), offering a compelling case for AI’s transformative power while also raising critical questions about its workforce impact.
Indications are that knowledge workers using ChatGPT are 37% faster and produce 40% higher quality work, but with a 19% decrease in accuracy. This suggests that AI can boost productivity and quality, but potentially at the expense of precision. However, it is optimistically proposed that user experience (UX) solutions could mitigate this issue.
A survey of enterprise users of Microsoft Copilot 365 further emphasizes the benefits of AI. The findings reveal that 73% of users agree that Copilot enhances their speed, 85% says it aids in faster drafting, and 72% report spending less mental effort on mundane or repetitive tasks. These results highlight AI’s potential to optimize workflows.
Interestingly, AI, particularly Language Learning Models (LLMs), benefits new or low-skilled workers the most, with a 43% improvement compared to a 17% improvement for more skilled workers. This implies that AI could be a potent tool for workforce upskilling and empowerment, serving as a support for early-career individuals while enhancing the toolkit of experienced professionals.
We need tools based on LLM (a type of AI) that can question our beliefs, help us assess situations, and offer different viewpoints. This means that while AI can enhance what we can do, it shouldn’t replace our ability to think critically and evaluate things. The idea is to have an AI that acts like Socrates, constantly encouraging workers to think more deeply and look at things from various angles.
A significant shift from content creation to content analysis and integration is posited. Future work may require more focus on understanding and interpreting AI-generated information rather than creating it. This underscores the growing importance of skills like critical thinking, leadership, and emotional intelligence – the human elements that AI cannot replicate.
AI can help dissect simple commands into micro-moments and microtasks, enhancing overall quality and efficiency. It is proposed that analyzing and integrating AI-generated information may become more crucial than searching and creating information, potentially altering the nature of work from content production to analysis and integration.
Interestingly, the binary narrative of “substitution vs. augmentation,” is challenged, proposing “innovation vs. automation” as a more nuanced framework. It is suggested that while AI may automate certain tasks, it can also catalyze entirely new roles and industries. The key is to monitor and foster this human-driven innovation alongside automation.
In conclusion, a compelling vision of an AI-augmented future is presented. However, it also emphasizes the need for critical evaluation, thoughtful implementation, and careful consideration of the ethical implications of AI in the workplace.
