Is AI A Useful Tool For Learning? Questions And Answers In the Style Of An Oracle.
Q: I got [AI software] to write me 50 sentences in the passato prossimo. Then I dumped them in a [AI language software] media file. Now I can practice them rapid fire with translations as well.
Will I learn Italian this way?
A: I can predict you will not speak Italian--or any foreign languages--at least not if you practice under the circumstances you have described.
Q: But AI can be a good and effective shortcut to quick and successful language learning, right?
A: When it comes to cognition, it is important to develop an efficient way of reaching specific goals. For language learning, a goal could be: to be able to use the passato prossimo and the imperfetto in combination automatically, effortlessly, without hesitation, and correctly in thirty days. Efficiency, however, does not imply shortcuts. The common misconception is that a shortcut cuts out the inessential. That is far from the truth. Learners only care to cut the work. They expect to finish a marathon in record time having never jogged before.
AI sells exactly this delusion. It writes the sentences for them so that they simply need to "dump" them into a different language software, which will automatically create a file, which they can open when they decide it's time to practice them "rapid fire". And it translates everything.
There's no happy ending to this plot.
Q: Ok, but the guy who practices with the 50 AI sentences will still learn something, won't he?
A: Efficient and successful acquisition and retention of information require independent physical and mental efforts combined. That guy will likely be able to memorize a few sentences, but he'll never be able to utilize the same pattern efficiently in an improvised conversation because he never trained to come up with his own sentences from scratch out loud.
Conversing requires interlocutors to produce their own words. Skip that in training and you won't be able to pull it off in real conversation where the stakes are much higher.
Q: Can we use AI to memorize vocabulary?
A: AI can give you a list of 100 vocabulary words on Gothic architecture. Useful if you memorize them and employ the proper techniques to retain them. Useless if you skim through the list once and expect to be able to use the words in conversation thereafter. Even with an efficient AI model (which doesn't exist), books already do that well enough.
Q: Can we use an AI model to have an actual conversation so that we can sharpen our conversation skills?
A: AI doesn't know when it should correct you and when it shouldn't. For now, it doesn't sound realistic, and this can greatly impact your pronunciation, especially in regards to prosody. AI will likely require you to always use perfectly grammatical sentences and might correct you when you're using an acceptable colloquialism, or a sentence that is ungrammatical by choice. Also, there are topics many AI models are not allowed to explore.
You can make do with it for lack of better options, but not in the long term.
Q: But AI can at least mimic an interaction with a real person. Isn't that better than repeating things or talking to yourself?
A: Yourself is a real person. AI isn't. Its replies are too perfect, too unemotional, too controlled, and too balanced. In no way does that remotely mimic a person-to-person interaction. Using AI also strips the experience of the high stakes and the emotional impact (stress, nervousness, anxiety, anticipation, excitement, fear, etc.) resulting from an actual interaction with a native speaker.
If you want a person-to-person interaction that doesn't just involve yourself, pick up the phone and book your next stay in Italy over the phone.
Q: But persons make mistakes all the time. How do I know if I'm speaking correctly or saying the right thing while conversing with a native speaker?
A: You make mistakes, don't speak correctly, and hardly ever say the right thing in your own language. I don't see why that's never an issue in your language, but it becomes unacceptable when you're trying to learn a new one.
Q: AI might make mistakes, be flawed, and overall be of little help. In your own words, AI is useless, but so are language teachers. So why is choosing AI over bad teachers such a wrong choice?
Q: Because with a bit of experience and common sense, you can easily tell whether a teacher is competent, but it's harder to detect AI's errors. AI doesn't hesitate, it wraps its answers in a veil of expertise, confidence, competence, and logical reasoning whether its answers are right or wrong, accurate or inaccurate, relevant or irrelevant.
Teachers at least offer human interaction, whether it's a positive or negative one.
In a brave new world where meaningful--and meaningless--human contacts are few and far between, bad teachers are still a better choice than AI.
Q: Can you name anything where you think AI can be useful?
A: Perhaps medical research, until it makes one mistake. Then you can't get justice, and insurance won't cover you because "AI doesn't make mistakes".
It also seems to be good for war. Then again, I don't see the usefulness in that.
Humans have survived thousands of years without AI, and the greatest geniuses that ever walked the planet didn't need AI.
Humans do not need AI.
Q: Do you use AI?
A: I don't, with one exception. I sometimes use it to create illustrations for my articles. Each article needs an image to accompany it, or no one will read it. But I never used AI--and never will--to do research, create content, write articles, edit them, develop my language system or any phase of the production of my videos.
Q: What do you use to write?
A: I write on my computer and with pen and paper in equal parts. I also often use a typewriter.
Q: Will there be an AI apocalypse?
A: I cannot imagine an apocalypse in the sense that humans will need to bunker down to survive a war against the machines--though governments and corporations have been stupid enough to actually make that possible.
Cognitively, however, we're in for some grim scenarios. The internet had already decreased our ability to learn and retain new data and skills by making information too easily and immediately available. AI has sped that process up to a dangerous paroxysm. People have lost problem-solving skills dramatically since AI was introduced. They have become fully dependent on AI, competence levels in all fields have decreased dramatically, and learning capabilities have become extremely reduced.
Q: AI can translate and interpret speech, text, linguistic signs, and communicative codes instantly. Doesn't that bring people together and create more unity? Doesn't that unite us by bridging the language gaps?
A: It surely could bridge certain language gaps, especially where communication needs to be practical and fast. However, communication needs to be practical and fast more rarely than we're led to believe. This being said, communicative gaps can also be filled in using verbal and paraverbal skills that are more or less universal and that humans are--or used to be--capable of handling naturally and instinctively. Individual creativity can kick in to help, adding human modes of expression to an otherwise sterile, hollow, and dehumanized type of interactions.
Q: AI can at least help us understand everything. Will you concede this is an important step forward in the development of communication and verbal skills?
A: It is important to fully comprehend a dinner order when waiting on patrons at a restaurant, but this requires basic, practical, and day-to-day verbal skills that don't require any AI intervention.
The idea that we need to understand every jot and tittle of every verbal exchange--let alone books, poems, music, works of art, concepts, philosophies, etc.--is unrealistic, fallacious, and impractical. And frankly, quite undesirable to boot.
A great degree of fascination for something arises out of the very fact that we don't understand all of it. Understanding--whether fully or partially--through personal research, comparisons, intuition, and individual mental/physical effort adds value to the word, the work, the art, the figure, the persona. But understanding by way of someone else's suggestions and interpretations makes it sterile and temporary. It also depletes the object of inquiry of a great degree of intrinsic value.
If the beauty of poetry can hardly survive the analysis of a bored and passionless university professor, one need only imagine how easily can AI destroy the value of something by cracking the entire code for us.
We appreciate something exactly because we don't understand all of it, and some codes aren't to be cracked.
Q: Is AI useful or effective anywhere with regards to learning a language?
A: To learning anything--I doubt it is.
To speaking a language, it surely isn't.
Q: Is AI in general a good thing for humans?
A: No.
AI is the worst news since smartphones and the advent of digital technology.
Comments