Political Affiliation and Attitudes towards AI

At a Glance

Research Question:

Do political views predict attitudes towards Artificial Intelligence?

Tech stack:

R Studio

Timeframe:

3 months

Size and source of dataset:

Survey data taken from American Trends Panel from Pew Research Center, over 10,000 responses with over 200 dependent variables.

Key methods or models:

Regression analysis

Headline results:

Individuals who self-identified as liberal showed more favorable attitudes towards AI.

Access:

Kaggle Notebook

Hypothesis Formation

Given my interest in technology and AI, I was interested in examining attitudes towards these kinds of technology and seeing where political affiliation was associated in anyway. My hypothoses for this project was: Political Affiliation (being liberal) has a positive effect on attitudes towards AI.

Data Collection and Cleaning

The data I imported from the American Trends Data was enormous with over 10,000 responses, and over 200 variables. The first thing I did was to examine the variables and pick out the ones that would help me answer this research question. With all of the variables related to AI isolated, I grouped them to better understand any potential nuances in the data. For example, some questions asked about an individual's concern for AI as opposed to their excitement towards it's development. This allowed me to see whether there were differences in an individual's concerns in addition to general attitudes.

The poll also included data on individuals' education level, as well as how much time they spend on the internet, allowing more in-depth analysis on people's behaviours in addition to their political beliefs.

Data Analysis and Interpretation

For my initial analysis, I performed a regression analysis and found that liberals were more likely to hold favorable attitudes towards AI. These findings were also shared among college-educated individuals, as well as those who spend more time using the internet.

Additionally, I found that college-educated individuals were less likely to think that AI was capable of making fair decisions