In that initial data set, I also found something else that I wasn't looking for: when it comes to attitudes about AI, where you live matters. I mentioned I'd dive into this further in a subsequent week. Here we are, in a subsequent week!
For this week's case study, I'm looking at how people in red states, blue states, and purple states feel about AI.
Nothing is bipartisan, but AI is close
Let's be honest, it's hard to think of many issues where Americans generally agree with one another. Covid, vaccines, and public health? Nope. Guns? Big nope. How to handle systemic racism, women's rights, and protections for trans people? Super big and emphatic nope, nope, and more nope.
But for most people, AI is new enough that they don't know what they think yet. Last year, I was invited to provide a private briefing for the Congressional Caucus for AI. I was pleasantly surprised by how open, curious, and bipartisan the conversation felt. It was just a bunch of curious people genuinely trying to understand very complex topics, and working together to do so.
Politically speaking, AI may still be up for grabs. But with the US being regionally divided on nearly everything else, I was only a little surprised to see some regional division in my initial data, and I wanted to understand it further.
To dive into this more deeply, I:
Picked 100 people with at least 50,000 followers who are posting about AI on LinkedIn at least 2x / month
Optimized for variation in US geography so I would get good geographic distribution, including representation from all 50 states
Sentiment-classified all their AI posts going back to January 1, 2024, which was about 1,400 posts in all
Classified all their home states as blue, red, or purple using the last three presidential elections. If a state always went Democrat, I called it blue; if a state always went Republican, I called it red; and if a state went different directions in different elections, I called it purple.
I validated the patterns I saw in the prior case study! Data is cool.
Who loves AI? Blue loves AI!
Obvious-but-important caveat: You can't tell someone's political orientation just from where they live. Unless someone's profile explicitly says MAGA4EVA, AMIRITE? or I BRAKE FOR TURTLES AND REPRODUCTIVE RIGHTS!!!, trying to discern political leanings at the individual level based on someone's LinkedIn posts is fraught with peril. I didn't even try.
That being said, posters who live in blue states are much more positive on AI than people living in red and purple states. Only 23% of the blue state posts have negative sentiment. compared to 39% of the purples and 42% of the reds. Almost half (46%) of the blue state posts have positive sentiment, compared to just 23% of the purples and 21% of the reds.
This purple is more like wine than periwinkle
When it comes to AI sentiment, purple is more like burgundy and less like indigo; in other words, purple staters are much closer to red than blue.
AI, here's where you're dead to me
I wondered if the red/blue distinction would be even more amplified if I focused in specifically on the AI posts that relate to government, politics, or elections. After all, does it even count as a modern election if it doesn't include AI-generated deep fake content designed for vital outrage?
Red and purple staters are equally skeptical about AI whether they're applying it to government, elections, or any other topic. Not so for blue staters: Turns out that despite their overall bullishness on AI, when it comes down to politics, blue staters are kinda leery. Check out how the blue staters change POV when we focus in just on the posts relating to government, politics, or elections:
It's all AI fun and games, until you're running for office
In other words, when it comes to AI and politics, everyone seems equally skeptical.
Robots, take heed
It's true that blue staters are more positive on AI as a whole than the reds and the purples, and the difference is significant. But it's also true that the difference between red and blue is less extreme than it is on other issues. If you did the same sentiment analysis for posts on abortion or gun rights, these groups are much less likely to cluster closely together.
Where a person lives is an imperfect proxy for their politics, but I feel validated in my initial sense that AI still has at least the potential to drive more partisan alignment than we're using to seeing on other issues. (Ask me again in the fall when we're in the thick of election season. Sigh.)
An unsolvable math problem Many years ago, Randi Zuckerberg famously said, "Work, sleep, family, fitness, or friends. Pick three." Her point was that, in a 24-hour day and a 7-day week, there's only so much time to go around. The laws of physics prevent one person from focusing on all of these things at the same time. Before becoming a parent 17 years ago, I optimized for work, fitness, and friends. Since becoming a parent, I've optimized for work, fitness, and family. And here's what I've...
How to make it to the NHL: Be born in January If you haven't read the original Outliers work by Malcolm Gladwell, it's interesting. Professional hockey players are overwhelmingly born in January, February, and March. How come? Most NHL players are originally from Canada. Canadian youth hockey leagues use a January 1 birth-date cutoff. A child born in January is nearly a full year older than a teammate born in December of the same calendar year. At age 8 or 9, a 10-to-11-month age gap means...
Good news, sexism is solved! If you didn't see the tech news last week, the annual Grace Hopper conference was cancelled this year. Last year for its 25th anniversary, over 25,000 people attended, mostly women early in their tech careers looking to build community. This year, it was cancelled with just a few weeks of notice. Of course, the cancellation makes total sense, because obviously the tech industry is at total gender parity now. Every engineering group is filled with women! Female...