Which Economic Tasks are Performed with AI? Evidence from Millions of Claude Conversations

This research report, 《Which Economic Tasks are Performed with AI? Evidence from Millions of Claude Conversations》, authored by the Anthropic team, presents a systematic analysis of AI usage patterns in economic tasks by leveraging privacy-preserving data from millions of conversations on Claude.ai. The study aims to provide empirical insights into how AI is integrated into different occupational tasks and its impact on the labor market.

Research Background and Objectives

The rapid advancement of artificial intelligence (AI) has profound implications for the labor market. However, systematic empirical research on AI’s actual application in economic tasks remains scarce. This study introduces a novel framework that maps over four million conversations on Claude.ai to occupational categories from the U.S. Department of Labor’s O*NET database, identifying AI usage patterns and its impact on various professions. The research objectives include:

1. Measuring the scope of AI adoption in economic tasks, identifying which tasks and professions are most affected by AI.

2. Quantifying the depth of AI usage within occupations, assessing the extent of AI penetration in different job roles.

3. Evaluating AI’s application in different occupational skills, identifying the cognitive and technical skills where AI is most frequently utilized.

4. Analyzing the correlation between AI adoption, wage levels, and barriers to entry, determining whether AI usage aligns with occupational salaries and skill requirements.

5. Differentiating AI’s role in automation versus augmentation, assessing whether AI primarily functions as an automation tool or an augmentation assistant enhancing human productivity.

Key Research Findings

1. AI Usage is Predominantly Concentrated in Software Development and Writing Tasks

• The most frequently AI-assisted tasks include software engineering (e.g., software development, data science, IT services) and writing (e.g., technical writing, content editing, marketing copywriting), together accounting for nearly 50% of total AI usage.

• Approximately 36% of occupations incorporate AI for at least 25% of their tasks, indicating AI’s early-stage integration into diverse industry roles.

• Occupations requiring physical interaction (e.g., anesthesiologists, construction workers) exhibit minimal AI usage, suggesting that AI’s influence remains primarily within cognitive and text-processing domains.

2. Quantifying the Depth of AI Integration Within Occupations

• Only 4% of occupations utilize AI for over 75% of their tasks, indicating deep AI integration in select job roles.

• 36% of occupations leverage AI for at least 25% of tasks, signifying AI’s expanding role in various professional task portfolios, though full-scale adoption is still limited.

3. AI Excels in Tasks Requiring Cognitive Skills

• AI is most frequently employed for tasks that demand reading comprehension, writing, and critical thinking, while tasks requiring installation, equipment maintenance, negotiation, and management see lower AI usage.

• This pattern underscores AI’s suitability as a cognitive augmentation tool rather than a substitute for physically intensive or highly interpersonal tasks.

4. Correlation Between AI Usage, Wage Levels, and Barriers to Entry

• Wage Levels: AI adoption peaks in mid-to-high-income professions (upper quartile), such as software development and data analysis. However, very high-income (e.g., physicians) and low-income (e.g., restaurant workers) occupations exhibit lower AI usage, possibly due to:

  • High-income roles often requiring highly specialized expertise that AI cannot yet fully replace.

  • Low-income roles frequently involving significant physical tasks that are less suited for AI automation.

• Barriers to Entry: AI is most frequently used in occupations requiring a bachelor’s degree or higher (Job Zone 4), whereas occupations with the lowest (Job Zone 1) or highest (Job Zone 5) education requirements exhibit lower AI usage. This suggests that AI is particularly effective in knowledge-intensive, mid-tier skill professions.

5. AI’s Dual Role in Automation and Augmentation

• AI usage can be categorized into:

  • Automation (43%): AI directly executes tasks with minimal human intervention, such as document formatting, marketing copywriting, and code debugging.

  • Augmentation (57%): AI collaborates with users in refining outputs, optimizing code, and learning new concepts.

• The findings indicate that in most professions, AI is utilized for both automation (reducing human effort) and augmentation (enhancing productivity), reinforcing AI’s complementary role in the workforce.

Research Methodology

This study employs the Clio system (Tamkin et al., 2024) to classify and analyze Claude.ai’s vast conversation data, mapping it to O*NET’s occupational categories. The research follows these key steps:

1. Data Collection:

   • AI usage data from December 2024 to January 2025, encompassing one million interactions from both free and paid Claude.ai users.

   • Data was analyzed with strict privacy protection measures, excluding interactions from enterprise customers (API, team, or enterprise users).

2. Task Classification:

   • O*NET’s 20,000 occupational tasks serve as the foundation for mapping AI interactions.

   • A hierarchical classification model was applied to match AI interactions with occupational categories and specific tasks.

3. Skills Analysis:

   • The study mapped AI conversations to 35 occupational skills from O*NET.

   • Special attention was given to AI’s role in complex problem-solving, system analysis, technical design, and time management.

4. Automation vs. Augmentation Analysis:

   • AI interactions were classified into five collaboration modes:

     • Automation Modes: Directive execution, feedback-driven corrections.

     • Augmentation Modes: Task iteration, knowledge learning, validation.

   • Findings indicate a near 1:1 split between automation and augmentation, highlighting AI’s varied applications across different tasks.

Policy and Economic Implications

1. Comparing Predictions with Empirical Findings

• The research findings validate some prior AI impact predictions while challenging others:

  • Webb (2019) predicted AI’s most significant impact in high-income occupations; however, this study found that mid-to-high-income professions exhibit the highest AI adoption, while very high-income professions (e.g., doctors) remain less affected.

  • Eloundou et al. (2023) forecasted that 80% of occupations would see at least 10% of tasks impacted by AI. This study’s empirical data shows that approximately 57% of occupations currently use AI for at least 10% of their tasks, slightly below prior projections but aligned with expected trends.

2. AI’s Long-Term Impact on Occupations

• AI’s role in augmenting rather than replacing human work suggests that most occupations will evolve rather than disappear.

• Policy recommendations:

  • Monitor AI-driven workforce shifts to identify which occupations benefit and which face displacement risks.

  • Adapt education and workforce training programs to ensure workers develop AI collaboration skills rather than being displaced by automation.

Conclusion

This research systematically analyzes over four million Claude.ai conversations to assess AI’s integration into economic tasks, revealing:

• AI is primarily applied in software development, writing, and data analysis tasks.

• AI adoption is widespread but not universal, with 36% of occupations utilizing AI for at least 25% of tasks.

• AI usage exhibits a balanced distribution between automation (43%) and augmentation (57%).

• Mid-to-high-income occupations requiring a bachelor’s degree show the highest AI adoption, while low-income and elite specialized professions remain less affected.

As AI technologies continue to evolve, their role in the economy will keep expanding. Policymakers, businesses, and educators must proactively leverage AI’s benefits while mitigating risks, ensuring AI serves as an enabler of productivity and workforce transformation.

Related Topic

HaxiTAG Intelligent Application Middle Platform: A Technical Paradigm of AI Intelligence and Data Collaboration
RAG: A New Dimension for LLM's Knowledge Application
HaxiTAG Path to Exploring Generative AI: From Purpose to Successful Deployment
The New Era of AI-Driven Innovation
Unlocking the Power of Human-AI Collaboration: A New Paradigm for Efficiency and Growth
Large Language Models (LLMs) Driven Generative AI (GenAI): Redefining the Future of Intelligent Revolution
LLMs and GenAI in the HaxiTAG Framework: The Power of Transformation
Application Practices of LLMs and GenAI in Industry Scenarios and Personal Productivity Enhancement

Read more →

SEO/SEM Application Scenarios Based on LLM and Generative AI: Leading a New Era in Digital Marketing

With the rapid development of Large Language Models (LLMs) and Generative Artificial Intelligence (Generative AI), the fields of SEO and SEM are undergoing revolutionary changes. By leveraging deep natural language understanding and generation capabilities, these technologies are demonstrating unprecedented potential in SEO/SEM practices. This article delves into the application scenarios of LLM and Generative AI in SEO/SEM, providing detailed scenario descriptions to help readers better understand their practical applications and the value they bring.

Core Values and Innovations

1. Intelligent SEO Evaluation Scenario
   Imagine a company's website undergoing regular SEO health checks. Traditional SEO analysis might require manual page-by-page checks or rely on tools that generate basic reports based on rigid rules. With LLM, the system can read the natural language content of web pages, understand their semantic structure, and automatically assess SEO-friendliness using customized prompts. Generative AI can then produce detailed and structured evaluation reports, highlighting keyword usage, content quality, page structure optimization opportunities, and specific improvement suggestions. For example, if a webpage has uneven keyword distribution, the system might suggest, "The frequency of the target keyword appearing in the first paragraph is too low. It is recommended to increase the keyword's presence in the opening content to improve search engine crawl efficiency." Such detailed advice helps SEO teams make effective adjustments in the shortest possible time.

2. Competitor Analysis and Differentiation Strategy
   When planning SEO strategies, companies often need to understand their competitors' strengths and weaknesses. With LLM and Generative AI, the system can quickly extract content from competitors' websites, perform semantic analysis, and compare it with the company's own content. Based on the analysis, the system generates a detailed report, highlighting the strengths and weaknesses of competitors in terms of keyword coverage, content depth, user experience, and offers targeted optimization suggestions. For instance, the system might find that a competitor has extensive high-quality content in the "green energy" sector, while the company's content in this area is relatively weak. The system would then recommend increasing the production of such content and suggest potential topics, such as "Future Trends in Green Energy" and "Latest Advances in Green Energy Technologies."

3. Personalized Content Generation
   In content marketing, efficiently producing high-quality content has always been a challenge. Through LLM's semantic understanding and Generative AI's generation capabilities, the system can automatically generate content that meets SEO requirements and has a high degree of originality based on the company's business themes and SEO best practices. This content not only improves search engine rankings but also precisely meets the needs of the target audience. For example, the system can automatically generate an article on "The Application of Artificial Intelligence in Healthcare" based on user-input keywords and target audience characteristics. This article would not only cover the latest industry developments but also, through in-depth content analysis, address the key pain points and needs of the target audience, significantly enhancing the article's appeal and utility.

4. User Profiling and Precision Marketing
   In digital marketing, understanding user behavior and devising precision marketing strategies are key to improving conversion rates. By analyzing vast amounts of user behavior data, LLM can build detailed user profiles and provide personalized SEO and SEM optimization suggestions based on these profiles. The system generates a detailed user analysis report based on users' search history, click behavior, and social media interactions, supporting the development of precise traffic acquisition strategies. For example, the system might identify that a particular user group is especially interested in "smart home" products and frequently searches for content related to "home automation" and "smart appliances." Based on this, the system would recommend that the company increase the production of such content and place related keywords in SEM ads to attract more users of this type.

5. Comprehensive Link Strategy Optimization
   Link strategy is an important component of SEO optimization. With LLM's unified semantic understanding model, the system can intelligently analyze the structure of internal and external links on a website and provide optimization suggestions. For instance, the system can analyze the distribution of internal links, identify whether there are unreasonable link structures between pages, and suggest improvements. The system also evaluates the quality and quantity of external links, recommending which external links need strengthening or adjustment. The system might point out, "A high-value content page has too few internal links, and it is recommended to increase the number of internal links to this page to enhance its weight." Additionally, the system might suggest strengthening cooperation with certain high-quality external websites to improve the overall SEO effectiveness of the site.

6. Automated SEM Strategy Design
   In SEM ad placement, selecting the right keywords and devising effective placement strategies are crucial. By analyzing market keyword trends, competition levels, and user intent, the system can automatically generate SEM placement strategies. The generated strategies will include suggested keyword lists, budget allocation, ad copy suggestions, and regular real-time data analysis reports to help companies continuously optimize ad performance. For example, the system might discover that "certain long-tail keywords have lower competition but higher potential conversion rates, and it is recommended to increase the placement of these keywords." The system would also track the performance of the ads in real-time, providing adjustment suggestions, such as "reduce budget allocation for certain low-conversion keywords to improve overall ROI."

Practical Application Scenarios and Functional Value

1. SEO-Friendliness Evaluation: By fine-tuning prompts, the system can perform SEO evaluations for different types of pages (e.g., blog posts, product pages) and generate detailed reports to help companies identify areas for improvement.

2. Competitor Website Analysis: The system can evaluate not only the company's website but also analyze major competitors' websites and generate comparison reports to help the company formulate differentiated SEO strategies.

3. Content Optimization Suggestions: Based on SEO best practices, the system can provide suggestions for keyword optimization, content layout adjustments, and more to ensure content is not only search engine friendly but also improves user experience.

4. Batch Content Generation: The system can handle large volumes of content needs, automatically generating SEO-friendly articles while ensuring content coherence and relevance, thus improving content production efficiency.

5. Data Tracking and Optimization Strategies: The system can track a website's SEO and SEM data in real time and provide optimization suggestions based on data changes, helping companies maintain a competitive edge.

6. User Behavior Analysis and Traffic Strategy: Through detailed user profiling, the system can help companies better understand user needs and adjust SEO and SEM strategies accordingly to improve conversion rates.

7. Link Strategy Optimization: The system can assist in optimizing internal links and, by analyzing external link data, provide suggestions for building external links to enhance the overall SEO effectiveness of the website.

8. SEM Placement Optimization: Through real-time market analysis and ad performance tracking, the system can continuously optimize SEM strategies, helping companies maximize the effectiveness of their ad placements.

Conclusion

The SEO/SEM application scenarios based on LLM and Generative AI provide companies with new optimization pathways. From evaluation to content generation, user analysis, and link strategy optimization, LLM and Generative AI are reshaping SEO and SEM practices. As these technologies mature, companies will encounter more innovation and opportunities in digital marketing, achieving more efficient and precise marketing results.

Related Topic

Research and Business Growth of Large Language Models (LLMs) and Generative Artificial Intelligence (GenAI) in Industry Applications - HaxiTAG
Enhancing Business Online Presence with Large Language Models (LLM) and Generative AI (GenAI) Technology - HaxiTAG
LLM and GenAI: The New Engines for Enterprise Application Software System Innovation - HaxiTAG
Leveraging LLM and GenAI Technologies to Establish Intelligent Enterprise Data Assets - HaxiTAG
Utilizing AI to Construct and Manage Affiliate Marketing Strategies: Applications of LLM and GenAI - GenAI USECASE
Leveraging LLM and GenAI: ChatGPT-Driven Intelligent Interview Record Analysis - GenAI USECASE
Exploring Generative AI: Redefining the Future of Business Applications - GenAI USECASE
Enterprise-Level LLMs and GenAI Application Development: Fine-Tuning vs. RAG Approach - HaxiTAG
How to Effectively Utilize Generative AI and Large-Scale Language Models from Scratch: A Practical Guide and Strategies - GenAI USECASE
Leveraging Large Language Models (LLMs) and Generative AI (GenAI) Technologies in Industrial Applications: Overcoming Three Key Challenges - HaxiTAG

Read more →

Cost and Quality Assessment Methods in AI Model Development

In HaxiTAG's project and product development, assessing the cost and quality of AI models is a crucial step to ensure project success. This process involves not only precise technical and data analysis but also the scientific application and continuous improvement of evaluation methods. The following are detailed steps for cost and quality assessment, designed to help readers understand the complexities of this process more clearly.

1. Define Assessment Objectives

The primary task of assessment is to clarify objectives. Main objectives typically include enhancing model performance and reducing costs, while secondary objectives may involve optimizing resource allocation and improving team efficiency. Quality definitions should align with key quality indicators (KQIs), such as model accuracy, recall, and F1 score, which will serve as benchmarks for evaluating quality.

2. Identify Cost Types

Classifying costs is crucial. Direct costs include hardware, software, and personnel expenses, while indirect costs cover training, maintenance, and other related expenses. Identifying all relevant costs helps in more accurate budgeting and cost control.

3. Establish Quality Metrics

Quantifying quality metrics is central to the assessment. Metrics such as accuracy, recall, and F1 score effectively measure model performance. By setting and monitoring these metrics, one can ensure the effectiveness and stability of the model in practical applications.

4. Conduct Cost-Benefit Analysis

Analyzing the cost-benefit of different quality levels helps identify the most cost-effective solutions. This analysis assists evaluators in choosing the best balance between quality and cost within limited resources.

5. Data Collection

Comprehensive data collection is foundational to the assessment. This includes historical data and forecast data to ensure that the assessment is supported by ample information for making informed decisions.

6. Cost Estimation

Estimating the costs required to achieve various quality levels is a key step. Estimates should include both one-time and ongoing costs to fully reflect the financial needs of the project.

7. Quality Evaluation

Evaluating the model’s quality through experiments, testing, and user feedback is essential. This phase helps identify issues and make adjustments, ensuring that the model’s performance meets expectations in real-world applications.

8. Develop Evaluation Models

Utilize statistical and mathematical models to analyze the relationship between cost and quality. Developing models helps identify the impact of different variables on cost and quality, providing quantitative decision support.

9. Sensitivity Analysis

Assess the sensitivity of cost and quality metrics to changes in key variables. This analysis aids in understanding how different factors affect model performance, ensuring the accuracy and reliability of the assessment.

10. Risk Assessment

Identify risk factors that may affect cost and quality and evaluate their likelihood and impact. This analysis provides a basis for risk management and helps in formulating mitigation strategies.

11. Decision Analysis

Use tools like decision trees and cost-benefit matrices to support decision-making. These tools help evaluators make informed choices in complex decision environments.

12. Define Assessment Standards

Determine acceptable quality standards and cost limits. Assessment standards should be set based on project requirements and market conditions to ensure the validity and practicality of the evaluation results.

13. Perform Cost-Quality Trade-Offs

Find the optimal balance between cost and quality. This process involves weighing the trade-offs between cost and quality to ensure effective resource utilization and achievement of project goals.

14. Implementation and Monitoring

Implement the selected solution and continuously monitor cost and quality. Ongoing monitoring and adjustments help maintain the desired quality levels and cost control throughout the project’s implementation.

15. Feedback Loop

Adjust assessment standards and methods based on implementation results. Feedback loops help refine the assessment process according to actual conditions, improving accuracy and practicality.

16. ROI Evaluation

Calculate the return on investment (ROI) to ensure that cost inputs lead to the anticipated quality improvements. ROI evaluation helps measure investment effectiveness and provides guidance for future investment decisions.

17. Continuous Improvement

Continuously optimize cost structures and enhance quality based on assessment results. Continuous improvement is crucial for achieving long-term project success.

18. Transparency and Communication

Ensure transparency in the assessment process and communicate results with all stakeholders. Effective communication helps gain support and feedback from various parties.

19. Compliance and Ethical Considerations

Ensure the assessment process complies with relevant regulations and ethical standards. This consideration is essential for maintaining the legality and integrity of the project.

20. Documentation

Document the assessment process and results to provide references for future evaluations. Detailed documentation aids in subsequent analysis and serves as a reference for similar projects.

In AI model development, assessing cost and quality requires in-depth expertise and meticulous data analysis. As technology evolves, assessment methods must be updated to adapt to new technologies and market conditions. Through scientific assessment methods, HaxiTAG can optimize project costs and quality, providing efficient AI solutions for clients.

Related Topic

Application of Artificial Intelligence in Investment Fraud and Preventive Strategies
AI Empowering Venture Capital: Best Practices for LLM and GenAI Applications
Exploring the Role of Copilot Mode in Project Management
Exploring the Role of Copilot Mode in Procurement and Supply Chain Management
The Digital Transformation of a Telecommunications Company with GenAI and LLM
Digital Labor and Generative AI: A New Era of Workforce Transformation
HaxiTAG Studio: Empowering SMEs with Industry-Specific AI Solutions

Read more →

The Path of AI Practice: Exploring the Wisdom from Theory to Application

In this new era known as the "Age of Artificial Intelligence," AI technology is penetrating every aspect of our lives at an unprecedented speed. However, for businesses and developers, transforming AI's theoretical advantages into practical applications remains a challenging topic. This article will delve into common issues and their solutions in AI enterprise applications, industrial applications, and product development, revealing the secrets of AI practice to the readers.

The Foundation of Intelligence: Methodological Choices

In the initial stage of AI product development, developers often face a crucial choice: should they use prompting, fine-tuning, pre-training, or retrieval-augmented generation (RAG)? This seemingly simple choice actually determines the success or failure of the entire project. Let's explore the essence of these methods together:

Prompting: This is the most direct method in AI applications. Imagine having a knowledgeable assistant who can provide the answers you need through clever questions. This method is ideal for rapid prototyping and cost-sensitive scenarios, making it perfect for small businesses and entrepreneurs.

Fine-Tuning: If prompting is akin to simply asking an AI questions, fine-tuning is about specialized training. It’s like turning a polymath into an expert in a specific field. For AI applications that need to excel in specific tasks, such as sentiment analysis or text classification, fine-tuning is the best choice.

Pre-Training: This is the most fundamental and important task in the AI field. It’s like building a vast knowledge base for AI, laying the foundation for various future applications. Although it is time-consuming and labor-intensive, it is a long-term strategy worth investing in for companies that need to build domain-specific models from scratch.

Retrieval-Augmented Generation (RAG): This is an elegant fusion of AI technologies. Imagine combining the retrieval capabilities of a library with the creative talents of a writer. RAG is precisely such a method, particularly suitable for complex tasks requiring high accuracy and deep contextual understanding, such as intelligent customer service or advanced Q&A systems.

Scientific Guidance: Implementing Methodologies

After choosing the appropriate method, how do we scientifically implement these methods? This requires us to follow a rigorous scientific methodology:

• Defining the Problem: This seemingly simple step is actually the most critical part of the entire process. As Einstein said, "If I had an hour to solve a problem, I'd spend 55 minutes defining it, and 5 minutes solving it."
• Conducting a Literature Review: Standing on the shoulders of giants allows us to see further. By studying previous work, we can avoid redundant efforts and glean valuable insights.
• Hypothesis Formation, Experiment Design, Data Collection, and Result Analysis: These steps form the core of scientific research. Throughout this process, we must remain objective and rigorous, continuously questioning and validating our hypotheses.
• Integrating Findings into the Existing Knowledge System and Sharing with Peers: The value of knowledge lies in its dissemination and application. Only through sharing can our research truly advance the AI field.

Practical Wisdom: Strategies and Steps

In actual operations, we need to follow a clear set of strategies and steps:

• Determining Metrics: Before starting, we need to define the success criteria of the project, which might be accuracy, recall rate, or other specific indicators.
• Understanding Constraints and Costs: Every project has its limitations and costs. We need to be clearly aware of these factors to make reasonable decisions.
• Gradually Exploring the Design Space: Starting from the simplest and most cost-effective solution, we gradually explore more complex solutions. This incremental approach helps us find the optimal balance.
• Tracking ROI: At every step, we need to evaluate the relationship between input and output. This is not only financial management but also a scientific attitude.

Challenges and Considerations: Core Issues and Constraints

In AI product development, we must also face some core challenges:

• Data Quality and Diversity: These are key factors influencing AI model performance. How to obtain high-quality, diverse data is a serious consideration for every AI project.
• Model Transparency and Interpretability: In fields such as medical diagnosis or financial risk control, we not only need accurate results but also an understanding of how the model arrives at these results.
• Cost and Resource Constraints: These are unavoidable factors in the real world. How to achieve maximum value with limited resources tests the wisdom of every developer.
• Technological Maturity: We need to consider the current technological level. Choosing methods that suit the current technological maturity can help us avoid unnecessary risks.

Conclusion: Co-creating the Future of AI

AI development is at an exciting stage. Every day, we witness new breakthroughs and experience new possibilities. However, we also face unprecedented challenges. How can we promote technological innovation while protecting privacy? How can we ensure AI development benefits all humanity rather than exacerbating inequality? These are questions we need to think about and solve together.

As practitioners in the AI field, we bear a significant responsibility. We must not only pursue technological progress but also consider the social impact of technology. Let us work together with a scientific attitude and humanistic care to create a beautiful future for AI.

In this era full of possibilities, everyone has the potential to be a force for change. Whether you are an experienced developer or a newcomer to the AI field, I hope this article provides you with some inspiration and guidance. Let us explore the vast ocean of AI together, grow through practice, and contribute to the human wisdom enterprise.

Related topic

Data Intelligence in the GenAI Era and HaxiTAG's Industry Applications
The Digital Transformation of a Telecommunications Company with GenAI and LLM
The Dual-Edged Sword of Generative AI: Harnessing Strengths and Acknowledging Limitations
Unleashing GenAI's Potential: Forging New Competitive Advantages in the Digital Era
Deciphering Generative AI (GenAI): Advantages, Limitations, and Its Application Path in Business
HaxiTAG: Innovating ESG and Intelligent Knowledge Management Solutions
Reinventing Tech Services: The Inevitable Revolution of Generative AI

Read more →