Microsoft(US:MSFT) 3 6 Ke·2025-07-16 11:13Core Insights - The study conducted by Microsoft Research analyzed 200,000 user interactions with Bing Copilot to assess the impact of AI on various professions, revealing that AI is not necessarily a threat to programmers but has significant implications for many other jobs [1][4][37] Group 1: AI Impact on Professions - The top six professions most affected by AI include interpreters and translators (98% of tasks can be performed by AI), customer service representatives, sales representatives, writers, technical writers, and data scientists [3][22] - Conversely, professions least affected by AI include nursing assistants, massage therapists, equipment operators, construction workers, and dishwashers, which require hands-on tasks [3][24] - The study found that high-paying jobs are not necessarily more susceptible to AI replacement, and that the applicability of AI varies significantly across different educational backgrounds [3][32][34] Group 2: User Interaction with AI - Users primarily engage with AI for tasks related to information gathering, writing, and customer communication, with AI often acting as a consultant or coach [3][12][21] - Approximately 40% of interactions showed a disconnect between user goals and AI actions, indicating that AI sometimes fails to meet user expectations [3][16] - The study highlighted that AI is more effective in assisting users rather than fully replacing them, particularly in knowledge-based tasks [21][34] Group 3: AI Applicability Scores - AI applicability scores were assigned to various professions based on the frequency of tasks performed by AI, completion rates, and the scope of AI's assistance [22][29] - The highest scores were found in knowledge-intensive roles, while manual labor jobs received the lowest scores, indicating a clear divide in AI's capabilities [23][24] - The analysis categorized professions into major groups, revealing that sales, computer and mathematical roles, and administrative support are where AI can be most effectively utilized [28][29] Group 4: Educational and Salary Correlation - The correlation between AI applicability and salary was found to be weak, with a correlation coefficient of only 0.07, suggesting that job type is a more significant factor than income level [32][34] - However, jobs requiring a bachelor's degree showed higher AI applicability scores, indicating that educational requirements do influence AI integration [32][33]