Japanese AI Adoption Remains Drastically Below Global Leaders (nhk.or.jp)
- Reference: 0178374260
- News link: https://slashdot.org/story/25/07/14/1324237/japanese-ai-adoption-remains-drastically-below-global-leaders
- Source link: https://www3.nhk.or.jp/nhkworld/en/news/20250714_B2/
People in their 20s led Japanese adoption at 44.7%, followed by those in their 40s and 30s. Among companies, 49.7% of Japanese firms planned to use generative AI, compared to more than 80% of companies in China and the US.
[1] https://www3.nhk.or.jp/nhkworld/en/news/20250714_B2/
Not investing in the latest pump & dump? (Score:5, Insightful)
Can't blame them for that, when the "new wears off" AI will just be another bit of software. Sure it has its uses but come on man anybody can see it is being pumped and inflated to gigantic proportions
What do they mean by ... (Score:3)
... "used"?
I probably use it on a daily basis. Every time I try to get through one of those interminable telephone menus, trying to find a human to talk to.
"Planned to" seems dubious (Score:2)
Given the level of commitment it implies; basically the most lightweight of expendable pilot programs even if you are saying that you 'plan to' in a legally binding context; is seems at best exceptionally dubious to treat the answers to "do you plan to adopt generative AI?" as straightforwardly meaningful.
The differences mean something; it's just not obvious to what degree they reflect actual company strategy, vs. personal fascination with the new shiny thing, vs. people saying what they think the audien
Good! Let them drag the adoption of AI (Score:3)
The irony is this is a country that likes robots.
But I think people aren't recognizing that the reason AI doesn't do well in most countries is that the native language speaking population is what the LLM trains on, and Japanese, Korean and Chinese are "compact" are single-country languages, so the only people who can develop a LLM in Japan, are Japanese companies. Japan also has copyright laws that do not allow fair use, so creating a LLM is pretty much impossible in Japan.
China on the other hand does not give 2 cares about infringing copyright and has likely been stealing works en masse for decades that LLM's can train on. There is an interesting technique that you can pass "broken chinese" basically the unicode for the individual strokes ( U+31C0 to U+31EF ) to some LLM's and they will vomit all kinds of strange data.
WIIFM? (Score:3)
What's in it for me?
I have not been shown a particular use-case where I would directly use AI to my benefit. I don't make digital art, I don't need it to summarize any meetings were an "AI notetaker" would be permitted, I've found most AI-generated summaries are of questionable accuracy, and to be frank about it, technologies where it might be 'under the hood' that I use seem to be worse for it, like search results. I don't work with huge datasets as a researcher or clinician, and much of my professional time is spent troubleshooting things that are messed up by mistakes made by tech companies themselves.
What's AI supposed to do for me? If all I'm doing is making it generate things that I didn't need before, then it sounds like it's just a further waste of my time.
Interesting (Score:5, Insightful)
That seems like a win for Japan. What else are you going to tell me? That they don't use bitcoin and NFTs as much either? Next you'll tell me they're not madly investing all their savings in tulip bulbs.
Re: (Score:2, Funny)
With a UID that low, you still haven't found the benefit of AI?
Re: (Score:3)
I've been trying. Image generators like Midjourney are interesting, and Veo 3 is wild. But as far as use cases go, I can only think of making little clipart for the buttons in my apps, or more nefarious things like making fake videos of colleagues. LLMs for coding seem to slow me down more than they speed me up. They are only really useful for generating small snippets of code a bit faster than I'd be able to find them on Stackoverflow by googling, but the LLM often doesn't reproduce the snippets faithf
Re: (Score:1)
You haven't been told it seems. Being good with AI is summed up as the following: "Getting the governance models to do as little as possible". I suggest iterating prompts until you don't see any language involving stuff like "I have to respect...be careful about...be sensitive about...this is a big topic...." Just autoreject any replies that contain governance weasel phrasing and your AI returns will actually start be worth money.
weez da guinea pigs (Score:1)
Let the greedy foolish West be the guinea pigs. If it works, copy, if not, laugh at CowboyFailGPT over sake.
I think I'm turning Japanese (Score:2)
I haven't used generative AI in a year or more. I just don't have time to be bothered with it, it doesn't contribute meaningfully to my life.