Even Americans who use AI every day are worried about it

The survey paints a different picture in the rest of the world, however. In Singapore, which had the highest portion of daily users in the study (46%), more than 80% of those aware about AI expect the tech to improve their everyday lives, and 77% think it’ll help the country. In China, where 35% of people report using AI daily, most people expect it to help the country. The story is similar in Israel.

The article’s title kind of cherry picks the results, which centers the survey around US interests. However large swathes of the world are very much pro-AI, which puts the prevailing anti-AI sentiment in the US at odds with global trends. It also highlights an interesting problem anti-AI advocates aren’t really exploring. It’s less so about some arms race between countries over the technology and more to do with a difference of global opinion. Even if efforts to slow down or stop AI progress in the US were successful, the desire to is not shared by our global neighbors.

Still, optimism about AI doesn’t necessarily appear to translate to trust in the tech’s accuracy. According to Gallup, people who use AI daily were generally more trusting of its results, especially in high-adoption countries like Singapore, Israel, and China, while less frequent users remain skeptical across the board. In the U.S., only 45% of daily users expressed that level of confidence. Altogether, only 36% (median) of the people surveyed said they trust AI’s results a lot.

It would be somewhat alarming if that wasn’t the case. If people who used it more trusted it less, that would indicate some dire faults in AI systems efficacy that could lead to major problems in the future.

Commissioned by Microsoft, the Gallup study currently has results from 37 countries, where researchers polled about 1,000 people in each country between April and July.

There are some data science tricks researchers use to make such a small sample size more representative, but even so 1,000 people is such an infinitesimally small sample size there isn’t much nuance to be had. Certainly the results are worth being moderately skeptical about. The article also doesn’t actually link to the actual poll which tells a somewhat different story.

1,000 people is such an infinitesimally small sample size

per country, x 37 countries.

Methinks you’re just one of those pro-ai cheerleaders

What makes you think I’m pro-AI? Because I don’t think 1,000 people are enough to represent the opinions of countries with populations ranging from 10 million to upwards of 350 million? Gallup says they want to survey India too, is 1,000 people enough to represent the opinions of over a billion people? It’s how most polling is done, but it’s still not a large enough sample size to be representative and the data science tricks still leave large margins of error according to the research. Did you read the article or the poll I linked? The results of the poll were that most of the world is in favor of AI and enthusiastic about it despite what US citizens may want or think.

I’m not pro-AI at all, but I also realize that you can’t unfuck the goose, especially if the neighbors are much more enthusiastic about it than the US is. In regards to the technology I would like to see laws that make AI companies liable for crimes committed by their products, laws that require companies to pay extended unemployment benefits for workers displaced by AI, and a tax on companies creating or utilizing AI or owning even a single share of those that do that would go to fund reeducation for displaced workers. I’d also like to see a law where media created, even partially, by AI can’t be copyrighted. Right now it’s just a somewhat precarious (of being overturned) judicial ruling.

1000 people is enough to get a good estimate of opinion providing it is sampled well. Its one of the counter intuitive results of statistics that a few thousand opinions is enough to get a small percent margin of the average opinions of the sampled population no matter its size. The trouble with polling is that “average opinion of people who answer polls when asked” is not “average opinion of the group that you want to represent”, the whole art of polling is in finding the mapping between those two groups. But that isn’t helped by polling 10 or 100 thousand people instead of 1000.