主頁/博客/AI Assistants Boost Beginners More Than Experts, Study Shows Correlation

AI Assistants Boost Beginners More Than Experts, Study Shows Correlation

There once was an AI named Chat who was really good at repeating back information it already knew. One day, Chat was given to some office workers [1] to help them with their jobs. Some of the workers were experts at their jobs, while others were still learning.

At first, Chat helped all the workers get more work done faster - even the experts! But soon, the experts noticed something funny. The workers who were still learning got way MORE help from Chat. The new workers improved a lot using Chat, doing their work faster and better than ever before!

The experts wondered why Chat didn't help them as much. That's when they realized - that Chat is an expert at repeating back facts but can't come up with brand new ideas. So, for workers who already knew those facts, Chat didn't offer them that much new help. But for newer workers still learning those basics, Chat was able to teach them so much more!

This shows a correlation - as in, two things that relate to each other and change together. The more expert a worker already was, the less helpful Chat was for them. But for newer workers, Chat could help them almost as much as the experts! It's because of their different starting points. Chat has a limit to how expert it can be. So, the closer a worker already was to Chat's expertise, the less new stuff Chat offered them.

The experts and newbies improved at different rates thanks to Chat. Their own expertise compared to Chat's matters for how much more they can learn. That connection in how much they improve is the correlation!

The SDTEST^® gives clues to someone's motivational values. However, additional polls can provide more pieces of the puzzle.

Imagine also giving an "A.I. and the end of civilization" poll. It asks people to rate at the agree or disagree level.

Now imagine 100 people who took both tests. You could match up each person's SDTEST^® colors with their rated answers about the danger of AI.

Comparing tests gives an expanded picture of values in action. More puzzle pieces make the whole image more apparent!

Multiple tests can work together, like colors blending on a palette. Other polls reveal what engages your values, like what is the perception of the danger of AI. Combined, they paint a richer picture of what motivates our thoughts and deeds.

Below you can read an abridged version of the results of our VUCA poll “A.I. and the end of civilization“. The full results of the poll are available for free in the FAQ section after login or registration.

人工智能和文明的终结

圖表相关性

VUCA

有一个统计上显著的相关性

正态分布，威廉·西莉·格塞特（William Sealy Gosset）（学生） r = 7.59

非正态分布，Spearman r = 0.31

仅显示结果 > r = 7.59

投票问题

所有问题

1) 安全（您同意或不同意多少？）

2) 控制（您同意或不同意多少？）

分配

正相关,%

负相关关系,%

普通的

1.8 Answer 1 1.2 Answer 2 5.2 Answer 3 1.1 Answer 4 2.7 Answer 5 #1	-6.4 Answer 2 -5.8 Answer 10 #1
1.2 Answer 6 4.3 Answer 7 2.2 Answer 8 8.6 Answer 9 #2	-5.3 Answer 11 -4.9 Answer 6 -3.2 Answer 12 #2

普通的

2.9 Answer 10 5.2 Answer 4 #1	-5.1 Answer 1 -0.2 Answer 2 -1 Answer 2 -0.9 Answer 3 -0.8 Answer 5 #1
4.8 Answer 6 1.1 Answer 12 7.8 Answer 8 0.5 Answer 9 #2	-7.3 Answer 11 -1.7 Answer 6 -2.1 Answer 7 #2

普通的

4.9 Answer 2 #1	-0.3 Answer 1 -0.8 Answer 2 -3.5 Answer 3 -0.1 Answer 10 -0.8 Answer 4 -0.2 Answer 5 #1
9 Answer 6 1.6 Answer 8 #2	-2.1 Answer 11 -1.2 Answer 6 -0.9 Answer 7 -8 Answer 12 -2.7 Answer 9 #2

普通的

4.9 Answer 2 4.8 Answer 10 7.9 Answer 4 #1	-10.6 Answer 1 -2.8 Answer 2 -4 Answer 3 -0.6 Answer 5 #1
5 Answer 11 3.1 Answer 6 3.5 Answer 12 0.3 Answer 8 #2	-0.1 Answer 6 -5.8 Answer 7 -8 Answer 9 #2

普通的

10.1 Answer 1 4 Answer 2 1.4 Answer 3 #1	-0.3 Answer 2 -1.7 Answer 10 -7.5 Answer 4 -5.5 Answer 5 #1
6.4 Answer 11 1.4 Answer 7 5.3 Answer 12 #2	-4.1 Answer 6 -3.9 Answer 6 -0.8 Answer 8 -3.2 Answer 9 #2

非正常

0.7 Answer 3 11.2 Answer 5 #1	-0.2 Answer 1 -0.3 Answer 2 -5 Answer 2 -0.9 Answer 10 -5.3 Answer 4 #1
2.2 Answer 11 1 Answer 7 3.1 Answer 12 6.1 Answer 9 #2	-2.9 Answer 6 -3.2 Answer 6 -4.4 Answer 8 #2

普通的

7.5 Answer 1 2.7 Answer 2 2.3 Answer 3 #1	-1.3 Answer 2 -0.3 Answer 10 -3.9 Answer 4 -6.2 Answer 5 #1
2.8 Answer 11 1.2 Answer 6 2 Answer 6 3.3 Answer 7 #2	-1.8 Answer 12 -11 Answer 8 #2

出口到MS Excel

您自己的VUCA民意调查中可以使用此功能

好的

[1] https://www.ft.com/content/b2928076-5c52-43e9-8872-08fda2aa2fcf

2023.11.27

Valerii Kosenko

产品负责人 SaaS SDTEST®

Valerii 于 1993 年获得社会教育心理学家资格，此后将他的知识应用于项目管理。
Valerii 于 2013 年获得硕士学位以及项目和项目经理资格。在硕士学习期间，他熟悉了项目路线图 (GPM Deutsche Gesellschaft für Projektmanagement e. V.) 和 Spiral Dynamics。
Valerii 是探讨 V.U.C.A. 不确定性的作者。使用螺旋动力学和心理学数理统计的概念，以及 38 个国际民意调查。

这个帖子有 0 评论

取消答复

发表评论