Anthropic Economix Index: 5 Key Insights

The Anthropic Economix Index was released on February 10, 2025.

Up until now, there have been a range of theories on how AI will be used across different industries.

However, the Anthropic Economix Index is different — it focuses on a specific, detailed analysis of AI usage, pulling data from millions of interactions with Claude.

As such, Anthropic notes that it offers “first-of-its-kind data and analysis” to provide insights into the impact of AI in today’s economy.

Here, we look at some of the key learnings from this index.

Key Takeaways (TL;DR)

1. 10.3% of the AI queries studied were for creative, media, or editorial tasks.
2. The majority of AI queries were for computer or math-related tasks (37.2%).
3. Although AI usage isn’t the same across all industries, about 36% of occupations used AI to some degree for around 25% of tasks.
4. Generative AI is predominantly being used as an assistant for human tasks, such as for brainstorming or validating work, rather than replacing human creativity.
   
   • Specifically Anthropic notes that 57% of users apply generative AI to augment or assist with tasks. In contrast, a moderately lower 43% of users incorporate generative AI for automation (where AI is conducting the work directly).
5. AI usage is most prevalent in mid-to-high-wage roles. They found that usage was lower for the lowest and highest-paid occupations.

Maintain transparency in the age of AI and review your content with the Originality.ai AI Checker.

5 Key Learnings From the Anthropic Economix Index

1. 10.3% of AI interactions were for creative and editorial tasks

While the data from the study shows that AI isn’t primarily used for automating work or jobs, it is being used consistently for specific tasks, including tasks that fall into the content marketing industry.

According to the findings, creative and editorial work (specifically defined in the report as “arts, design, sports, entertainment, and media”) was the second-highest usage category, making up 10.3% of all queries studied. 

Within this category, Anthropic further highighlighted that writing and editing, marketing and promotional strategies, and public relations were all popular tasks the tool was used for.

Looking for more insight on how to incorporate AI into content marketing? Learn more about brainstorming content with AI, content research, and competitor analysis.

2. The majority of AI queries were for computer or math-related tasks

However, content marketing related tasks such as writing and editing aren’t the only way generative AI is being incorporate into work.

“Computer & mathematical” queries such as software engineering or development and coding accounted for the highest level of use at a whopping 37.2% of the total usage.

This makes sense, as AI is a fantastic tool for modifying and tweaking code, as well as identifying issues that are difficult to see with the naked eye.

Software engineers are also more likely to feel comfortable with new technology like AI, considering technological skillsets and training, which would also account for why this industry makes up such a high percentage.

3. AI usage is present across multiple industries

However, while AI isn’t used at the same rate across all industries, it is nonetheless being incorporated into a range of applications.

Typically the report found that usage is significantly lower for industries with in-person interactions, such as transportation (0.3%) and healthcare support (0.3%) or healthcare practitioners (2.6%). 

These lower numbers are an apt reminder of the current limitations of generative AI.

While the depth of AI use varied from occupation to occupation, the general use of AI is notable according to Anthropic, with about 36% of occupations incorporating AI (at some degree) into a quarter (25%) of tasks.

4. Generative AI isn’t replacing human creativity

Whilst generative AI usage is on the rise (read our study on AI in Google Search Results), there are plenty of signs that users are implementing Claude as an assistant for their tasks rather than a replacer.

This collaboration is shown in the prompts, where users focus on learning, brainstorming, and double-checking their work.

“In just over half of cases, AI was not being used to replace people doing tasks, but instead worked with them” - Anthropic

5. The highest AI usage rates were in mid-to-high salary brackets

As part of the index, Anthropic also pulled in salary data from O*NET, which tracks the median U.S. salary by occupation.

As a result, the data shows that the highest use of AI came from mid-to-high salaries, and interestingly, the lowest and highest-paid occupations had a much lower level of uptake.

This could possibly be due to positions in the mid-to-high salary bracket covering managerial, admin, project management, and strategy roles, which can benefit from generative AI support.

In contrast, many lower or higher-paying jobs (such as medical practitioners) require a higher level of specialized manual or in-person work. 

Final Thoughts

While the Anthropic Economix Index is incredibly insightful, it does have some limitations to be aware of.

The most notable limitation Anthropic described in the study was that there wasn’t certainty around whether Claude was being used to complete a work task or a hobby project. They provided the example of writing advice, which could be used for work or for drafting a novel for fun.

There are also other caveats, including only looking at data from the Free and Pro plans, excluding Enterprise users, and that some of the tasks may have been classified incorrectly, due to the volume of data.

Keeping in mind the limitations of the study, the results still offer unique insights into how AI is changing the workplace and help paint a more detailed picture of true AI usage across industries, from software development to the arts and copywriting to healthcare.

Establish transparency in the content you publish with the best-in-class Originality.ai AI Checker.

Then, learn more about the latest developments in AI in our top guides:

• OpenAI Introduces Operator — Key Things to Know
• DeepSeek: What Is It and Why It’s Disrupting the AI Industry
• Gemini Statistics, Facts, and Product History