Hi all, I have been thinking of a business idea that applies both AI and education (2 of my favorite fields). What if I form a startup that develops AI English language translation services for online programs in Europe/Asia/etc. I was thinking it could be something like the ENEB idea where you can study a Spanish propio in English. The difference is my product will apply to any university and any language. The selling point to the universities, is that the AI translation abilities would allow them to market their degrees to the world. It is also a lot cheaper than hiring English interpreters or English speaking faculty. For lectures, the AI can produce English subtitles like what you see on foreign films on Netfix. The technology is there - I don't think anyone is using it widely in the education sector. It can also be sold to US/Canadian companies that want to hold classes on-campus to Chinese students who may not be that good or do not want to learn English (e.g. they intend to go back home to China). The selling point to the consumers is they get access to all universities in the world, regardless of language. I can be French and study at a Chinese university and vice-versa. The world is getting smaller everyday. We all have access to the same information/news/websites, why not have access to all education? That will also force US universities to compete with lower costs offered by foreign counterparts. It's a win-win for everyone. I don't mind if any of you steal my idea, but if you find this interesting, maybe we can all pool our minds together to come up with something mind-blowing.
I don't think that AI translation is quite there yet. It's better than it was 10 years ago, I'll grant you that, but it is hilariously wrong at times. Translation is kind of okay for French, Spanish, and Italian. I assume also Portuguese, but I don't go looking for Portuguese content. It's slightly worse for German. Bad for Slavic languages (Russian, Polish, Serbo-Croatian, etc.). And absolutely terrible for Asian languages. I assume the same holds true of African languages, except perhaps Afrikaans. I have had automatic translation give me the exact opposite meaning of what is being said for news stories, so I would be extremely skeptical of using it in an academic context. It switches genders of pronouns, even when translating from a heavily-gendered language. It adds "not" when it shouldn't be there and drops it when it is. DeepL is much better about this than Google Translate is, but they both have issues that would likely frustrate anyone attempting serious study.
This idea would be great, expensive to build your own AI system, but if you have the ability and means it would be awesome. A way cheaper and faster solution would be to set up a program in a state with the most liberal policies for doing so, approach foreign programs and tell them you'll translate their full curriculum into English and make a deal, set up an automated online exam system like practically every school has now (or make it entirely paper-based), and have a grading company take care of all of the paper grading like many schools are doing these days (and just not telling us): https://www.chronicle.com/article/some-papers-are-uploaded-to-bangalore-to-be-graded I would avoid a propio situation and try to get the schools to award the degrees themselves and handle the transcripts themselves. Some schools will do it like San Jorge University does.
100% agreed. I wouldn't dream of trying to do a degree program in a language I couldn't read or write.
I totally agree, although that is the whole point of the startup - to develop it to the level where it can be used in Academics. I am pretty sure we can get it to a good enough level to use for certain fields like computer science, data science, statistics where there is more emphasis on numbers. I agree it will probably take some time to get this applied to fields like philosophy, law etc where there are nuances in the usage of language. But even if the startup succeeds at even just one of those fields like computer programming, it will be a big win worldwide for all universities.
That's my goal - to come up with a product good enough to satisfy even the most passionate educators like Steve . Someone will have to figure out this technology eventually, may as well start it now and establish partnerships early on in the game. When the tech then matures, the startup would already have a foothold (first mover advantage) in the industry. To be honest, I think a company like Edx or Coursera would be interested in buying a startup like this. It is definitely the future of education I think. THE next step in the distance learning industry.
This is a great idea. However, I don't think it'll go in many markets. In Europe, especially the Nordic ones, English is a second language to many countries, it's like an immersion program, they're very fluent. The same goes with South/South East Asia, many within India, Malaysia, Pakistan, Philippines have English as a second language from school, this is because English is an official language of the country. Other parts of the world have this official language as well, such as countries in Africa. You will only have a very niche market, it may be hard to target, your only "vertical" to target are countries that do not have English as an official language, but the market is slim. An example that won't go well is China, sadly, they just eliminated foreign owned English tutoring last month...
People have been working on the issue of reliable automatic translation for a long, long time. It seems like a good idea in theory, but the reality is that I don't think that anyone is going to solve the problem in the next 3-5 years at least. DeepL is getting there, but it's still going to take some time.
Yes, but if you wait until the technology is developed before setting up the foundation (case uses) for universities, then you risk being the last guy/gal in once that technology evolves. Basically, the idea is I am absolutely certain the technology will get there. In my job, we are already using / applying some of this tech. I attend meetings in French and the AI automatically converts it to English subtitles. Is it 100% accurate? Of course not, but I pretty much understand what my French Canadian colleagues are saying 80-90% of the time. The powerpoint slides are accurate around 90% of the time. Again, the idea is to be ahead in terms of applying the technology and partnering with universities so that once the technology has been fully developed, you are first in line to benefit. If you wait for 5-10 years when this is developed, you will be last in line and there will be no one left to partner with at that point. But I love that you are challenging this idea. I am very open to learning those pitfalls. It will allow me to prepare answers to potential questions from investors when I pitch this.
LOL, then consider LA's idea, because that's workable today, and there's room for that in the marketplace, especially if you can make English-language programs accessible to those in Francophone Africa.
Well once you come out of boost strap mode and decide to raise venture capital or cash out, always keep in mind how some similar technology was handled... https://www.nytimes.com/2012/07/15/business/goldman-sachs-and-a-sale-gone-horribly-awry.html
I´d also love to develop an educational offer for a propio Bachelor in Business Administration, maybe a DBA. Buy some already existing course material, develop an automated exam system. Find a partner institution which awards the propios. Sell it on Groupon. If you are interested in working on the smartdegree let me know
Interesting. My University of Florida program had instructors from University of Edinburgh. Even though we speak the same language and live under common law systems, there were some dialect and legal differences that led to some frustrations.