GPT-4 Makes Data Analysis Work 1% the Cost of Humans

The cost of GPT-4, the latest version of LLM developed by startup OpenAI, is just 0.45% of that of hiring a senior data analyst — who earns about $90,000 annually, or 0.71% of that of a regular analyst — according to researchers from Chinese company Alibaba.

In this study, the researchers posed a question and provided data, and automated the entire data analysis process using GPT-4 to extract and analyze the data, then generate insights and charts. The analysis results were compared to professional human data analysts in terms of performance, time, and cost.

According to the study, the tests showed that GPT-4 was not only significantly cheaper than a human data analyst, but also completed the tasks much faster. However, the study noted that more research is needed before concluding that GPT-4 can replace data analysts.

In some cases, the AI ​​model outperformed human data analysts in terms of accuracy of metrics and analysis, while the insights GPT-4 generated tended to be more complex, according to the study.

However, GPT-4 fell far short of humans in terms of accurately displaying data in charts, as well as in presentation and formatting in some cases. The study found that despite some errors in some numbers, GPT-4 was still able to produce accurate analysis.

The study sheds light on the expanding application of LLM in various industries, where it is expected to improve efficiency but also threaten human jobs. Chinese tech companies are racing to develop their own ChatGPT-like services.

Alibaba is working on its own solution called Tongyi Qianwen, which launched in April and is now available to enterprise customers by invitation, weeks after Baidu announced its Ernie Bot service.

Mai Van (according to SCMP)