(Continued from Part 1)
Once they are paired with technology like 3D printers, AI will likely be able to single-handedly create objects both small and large such as furniture on the spot.
I am sure that when E42 was planning this sentence, he meant to write something like, ‘AI will likely help people to create …,’ which would have been a good translation. Somehow, however, ‘people’ got lost when putting words on the page, and AI became the creator. I am afraid to say this small slip was a costly mistake, because the idea that people create whereas AI models generate is so central to this article. If E42 had used DeepL or some other app for a retro-translation check, I am sure he would have identified the oversight immediately, and might also have noticed that ‘on the spot’ is not the best option for たちどころ here either. On the credit side, I think ‘single-handedly’ very effectively captured what this sentence was about (and proves that E42 really was thinking about people all along).
On a relatively minor point, I noticed that E42 switches between ‘we’, ‘they’, and ‘you’ as the subject in some of the early sentences. This is one area where journal sub-editors often expect (and enforce) consistency, which is a useful reminder that we should get a birds-eye view of our text now and again, and not get tied down in sentence-by-sentence compartments.
…, advances in AI technology will make it next to impossible for artists with average skill to compete, which in turn will negatively impact their ability to make a living.
E42’s translation got close to the original アーティストが職業としては成り立ちにくくなる, but I worry there is some divergence. The focus is more on ‘artist’ as a job title, rather than the people who call themselves artists, because some (perhaps many) of the people covered here will never be able to make it as an artist. For them, rather than negatively affecting their ability to make a living, AI will deny them that ability completely.
Even today, the proportion of jobs to applicants for workers like artists and designers is only approximately 0.18, so just under one in five people are able to secure employment. This figure will only lower more and more in the future.
E42 made a nice try to explain our artist-related employment statistic to readers, although I wonder if ‘proportion’ would have worked better with a percentage than a decimal. Also, changing ‘one in five people’ to ‘one in five applicants’, and ‘lower’ to ‘get lower’ might have brushed things up a little.
Not only artists but also most white-collar workers including office workers, certified public accountants, tax accountants and academics and researchers like me can’t escape the influence of generative AI.
E42 has got all the key points across, but readers have an awfully long wait for the main verb, escape. Sometimes such delays in stating the verb are unavoidable, but good writing practice in English often involves keeping the subject short. With a couple of small edits, we could make this translation so much easier for readers. Imagine the readability score for a passage in the following pattern: ‘Artists are far from the only ones who cannot escape … . The same is true of ….’ Also, I believe that for journal articles at this level, ‘myself’ would look much better than the informal ‘me.’
What the current AI lack is will, first-hand experience and judgement of value.
I thought ‘first-hand experience’ was a nice imaginative choice for 体験, and I spent a lot of time thinking about ‘judgment of value’ for 価値判断. I see E42 used ‘value judgement’ later in his translation, and I would be interested to know if he regards the two terms as being equivalent. On reflection (and with a second opinion from my colleagues), I think they are, in which case the more widely used and recognised ‘value judgment’ would be a better choice, so long as the translator is not worried about countability (see my general comments). Translators who do worry about this countability issue might want to think about other terms, such as ‘subjectivity’, or even proposing an action-describing sentence.
Although AI like ChatGPT have high deviation scores indicating great academic ability, they are also likened to humans that wait for instructions and don’t take the initiative.
I think ‘they are likened’ is a great way to signal figures of speech to readers, but this phrase might have been more effective if it had been used at the start of the sentence. E42 opted to explain 「スーパー偏差値エリート」 to readers here, but I fear the game will be lost as soon as readers see the word ‘deviation’. That word evokes such powerful, negative mental images, that readers will never focus on the concept of an entrance exam score deviating from the mean by many positive orders of magnitude. Why not try to get closer to the full (original) meaning of スーパー偏差値エリート? We could, for example, steal a nearly equivalent category used for Australian and New Zealand Certificates of Education and try ‘... has been likened to a Very High Academic Achiever ...’ .
…, generative AI are akin to shut-ins confined to their room …
E42 shows a nice way to introduce figures of speech again with ‘akin,’ but I must confess that the word ‘shut-in’ was new to me. I checked out is usage (which possibly tends to be North American?), and it seems often to be associated with physical or psychiatric disability. I certainly cannot object to this translation, but I worry that some readers might infer that AI is sick in some way.
Therefore, they are not capable of creating a work whereby several original ideas are brainstormed and the most preferable one is chosen.
I think ‘they are not capable of’ is a good choice to highlight the total inability of AI at this point. However, I can also see a couple of problems. Firstly, it seems 人間のごとく has slipped through the net. Secondly, the ‘whereby’ part still leaves some ambiguity: could AI perhaps still manage some small part of the brainstorming and choosing preferable idea process? A solution might be using two clauses, the first stating AI models are not capable of things people can do, and the second stating specifically that it cannot do the brainstorming and choosing part. As I commented to E05 and E20, I think ‘therefore’ risks tying readers to a one-paragraph interpretation of this sentence.
If you instruct an image-generative AI to draw Tokyo Tower in the style of Van Gogh, it will create such an image for you. However, it can’t create style in the first place.
I thought this simple translation was an excellent one; it did not get sidetracked by OO風, and rendered そもそもの in a way that was immediately intelligible to readers.
It is true that it is also difficult for humans to be original, for instance with traditional forms of media like pictures.
E42 has chosen to simplify this sentence, which is often a good strategy, but I feel it’s sad that we lose some of the atmosphere of an author-reader dialogue that the original had (in my subjective opinion). Furthermore, もはや looks to have been neglected here, but it is actually quite important to show there was once a time when people could be original in the medium of oil painting. I guess ‘that’ is intended to point back at the task of portraying Tokyo Tower (with originality), but I worry many readers will just take it as part of the ‘It is true …’ structure. How about ‘It is true now that this task would also be difficult for people/any of us….’?
Oil paint and cameras are forms of technology, and pictures and photographs are forms of media.
E42’s double use of ‘forms’ made for a nice translation. I am not sure exactly what the definitions of ‘technology’ and ‘medium’ would be in the field of art, but I think terms like ‘form’ here would give even the purists something they could live with.
From a different perspective though, as long as technology continues to advance, new art genres will continue to take form, so humans won’t run short of opportunities to demonstrate their innovative creativity.
E42 has covered all the points here, but I think our author might be a bit disappointed with ‘will not run short’ (does it sound a bit too much like something we would say before a shopping trip?).
Current Japanese education which drives children to cram knowledge into their heads and moulds them into point-grabbing machines will require drastic reform.
E42 deserves kudos for stating it is Japanese education that we are looking at. He was also the only one to translate 教育 as simple ‘education’ not ‘education system,’ which I think captures what the author is talking about nicely, and he leaves readers in no doubt of the requirement for drastic reform. All in all, this was a very nice, strong conclusion.
E46
E46 produced a highly accurate translation, all captured in very polished English. I get the impression that E46 enjoys playing with language, and that really helps with translations like this. Our text is a piece of non-fiction, but it is supposed to be of high literary quality, and there is a premium on clever use of English for such articles. My main suggestion for improvement would be that E46 should back his imagination more often. For example, in a few cases where he took ‘safe’ translation options, I felt he risked making an English sentence seem vaguer than the Japanese original would have been to its readers, by leaving the performer of an action unidentified. However, such overly safe choices are few and far between and do not detract from an excellent translation.
The Instant Idea-to-Product Economy and the AI Behind It
I thought the ‘AI Behind It’ part of the title worked well stylistically and seems to establish the right sort of relationship between the economy and AI, given the context of this article. However, even the hyphens do not make the idea-and-product part of the title catchy enough in my opinion.
ChatGPT, the text-generating AI tool making headlines of late, is capable of not only answering all manner of questions, but also generating essays, novels and computer programs as well.
As I commented elsewhere, squeezing the translations of 今話題 and 生成AI into a side comment early in the opening sentence gives readers neither a smooth sequence nor a fully coherent thread to the main discussion. Also, ‘making headlines of late’ seems like a rather dull 今話題. On the plus side, the translated sentence is certainly complete and accurate, and ‘all manner of questions’ is a really good choice for あらゆる質問.
On the other hand, image-generating AI like Stable Diffusion allows users to generate matching illustrations and photorealistic images simply by inputting phrases (such as “a couple frolicking at the beach”).
E46 brings people into the story at an earlier point in the text than any other finalist, with the ‘users’ here, and I have reason to believe that this is an inspired choice: Chapter One in the author’s latest book—published just one week after our contest’s entry date—is titled 「アイデア即プロダクト」の経済, and has several sentences with ユーザーas the 主語! However, I think we may run into a couple of problems with ‘matching’. Readers could easily assume the match here is between illustrations and photorealistic images, rather than between the AI model’s input and output. Even if readers avoid that pitfall, ‘match’ may not accurately describe what image-generative AI models can do. I tested this myself by getting Stable Diffusion to depict ‘a couple frolicking on the beach’ (via a free app), and the resulting image cannot really be described as a match; it seemed more as if Stable Diffusion had taken the input as a hint rather than a specification (try it yourself and see if you agree). I believe this is why the author chose 相応する to characterise this input-output relationship, which is less sensitive and selective than the 答えられる relationship he describes for ChatGPT.
Generative AI tools such as these can be used to produce picture books, manga, and more, which can then easily be sold as e-books. It is likely just a matter of time until users will be able to pair this technology with 3D printers and so forth to instantly create small items, furniture and more all on their own.
The first sentence here is a pretty good summation of what is going on the source text at this point; however, I think we should pause to consider that someone is making and selling these picture books, and that this someone (or some people) is understood by the writer and readers. E46 mentions ‘users’ in both the preceding and subsequent sentences, so the missing human actor in this sentence may not be a big problem. Even so, it is always worth asking ourselves ‘Who is the actor?’ whenever we see a subject-less active voice structure in a Japanese sentence. Failure to identify an actor has famously been identified by Jay Rubin as one of the greatest sins of Japanese-to-English translation!
Readers may be jarred by four instances of ‘and so forth’, ‘and more’, or ‘such as’, in just two sentences; I think we would be better off with reducing this count. Of course, など and や can be the bane of J>E translators, but when their influence on the English text is too obvious, readers will quickly pick on the unnatural expression, and suspect that the text is a translation. In addition to my worries about など, I have possibly the most pedantic comment in this year’s contest. I recommend E46 to look up the debate about the Oxford comma, which he did not use here (before ‘and more’), but he did use in other parts of the text. This may seem a ridiculous point but, in my experience, the Oxford comma issue can really exercise clients (and sub-editors), so we should try to aim for consistency on it (when we remember).
On the plus side, E46 uses ‘just a matter of time’ and ‘paired’ to make a very natural expression. This is exactly the sort of language that is a perfect fit for an opinion piece in a literary/current affairs journal.
AI technology opens up the door …
Here is another great example of imaginative use of English, which captures the original meaning while disguises the fact that this is a translation from unwitting readers. English has so many ways to say 可能にする, and E46 deserves a lot of credit for exploring those possibilities and finding something that fits the bill so well.
At the same time, this means that those with merely average skills will stand little chance against AI, limiting the viability of artist as a career choice. The current ratio of positions to applicants for active artist/designer job postings sits at 0.18 positions per applicant, meaning that only one in every five applicants will be able to find a job in the field.
This part of E46’s translation again shows natural and imaginative use of English. I think ‘limiting the viability of artist as a career choice’ was a really nice touch that captured アーティストが職業としては成り立ちにくくなる well, and adding ‘in the field’ ensures readers can make sense of the sentence. Maybe a bit more polishing is possible: if we put ‘one in five applicants’ up front, then maybe we will not need to write ‘positions’ and ‘applicants’ two times to explain the ratio. E46 has chosen to drop the ほど and ぐらい from the translation, but I think the ‘one in five’ figure at least should have ‘around’ in front of it (although maybe readers can guess that the 0.18 figure is rounded).
Artists are far from the only ones to be affected. From certification-bearing professionals such as office workers, accountants and tax specialists to educators and researchers (such as myself), the vast majority of white-collar positions will be unable to escape the reach of generative AI.
According to (my) rough-and-ready convention, good English writing involves a mix of shorter and longer sentences, with shorter sentences favoured at the start of a paragraph. I thus applaud E46 for breaking the original long Japanese sentence into two English sentences, and I think readers would subconsciously applaud too. However, I would be interested to ask E46 why he chose ‘certification-bearing’ instead of something like ‘certified’; either way, I think we should move ‘certification’ so it applies to the accountants and tax specialists only.
… this certainly doesn’t mean that AI will trump human workers in all regards.
The original 全での面で人間を上回る is quite hard to capture naturally in English, but I think E46 does a great job, again showing how a bit of imaginative usage can really bring the language alive (something I believe the author does well in Japanese).
Current AI lacks three elements: volition, experience, and judgment.
On a minor point, the colon is a nice example of how we can replicate the effect of the 鉤括弧in the original, while following the rules of punctuation for English.
I think ‘lacks’ is OK, but I worry a little that ‘lack of experience/judgment’ may not mean the same as ‘total absence’ to some readers.
In my general comments, I mentioned that ‘value judgement’ may not be the ideal term for an uncountable property, and I wonder if E46 agrees with me, given his choice of ‘judgment’. There is certainly a case for using ‘judgment’ alone here, but could it be a bit too wide in meaning? After all, AI models make some sort of algorithmic judgments in their processing, and I believe the author is specifically looking at subjective judgements here. Given the context, our English readers may deduce all this correctly, but it is always a little risky to rely too much on context in English!
ChatGPT is … the crème de la crème of the intellectual elite
I thought E46 did exceptionally well to find an idiom that matches the author’s idiomatic phrase in Japanese. To make the translation more accurate, I would suggest replacing the overly sociological ‘of the intellectual elite’ with just ‘academically’. But we may have a problem: I just don’t think our author will go for ‘crème de la crème’. Remember, we are writing for a borderline Generation X/Millennial economist with a background in IT who very much wants his prose to be accessible to the younger generation. Much as I liked ‘crème de la crème’, I am not confident that an idiom derived from French fits an authentic voice for this author. So, in real life, we might have to come up with a Plan B; perhaps a simpler ‘top one per cent academically’ would fit the bill.
Generative AI can also be likened to a “super shut-in” of sorts, cooped up in its room and completely absorbed in reading or browsing the web. It has never experienced a campfire or digging for shellfish when the tide is low, nor will it ever produce a novel or movie based on its own experiences.
I was surprised E46 reverted to quote marks here, although maybe they can de defended on the grounds that ‘shut-in’ is an unfamiliar phrase. As I have commented elsewhere, I think ‘super’ is a slightly unnatural choice in written English here; maybe ‘the ultimate’ or ‘a total’ would be more natural.
The (big) credit points here come from the ‘has never …. nor will it’, which worked wonderfully well for ことがなく、…. こともない. This is a great illustration that translation is about more then word choices, and there is much we can do with sentence structure to convey meaning. Also, ‘cooped up’ is yet another great example of E46’s imaginative use of English.
Moreover, AI is unable to make judgments on aesthetics or issues of right and wrong, falling short of anything more than merely imitating human decision making.
The first half of this sentence was a very nice translation, especially the choice of ‘aesthetics’ for 美醜. However, a tough journal sub-editor might get his red pen out for ‘falling short of anything more than …’, and honestly, this wording could easily confuse readers: exactly what level of judgement was AI supposed to be aiming for? Something simpler on the lines of ‘doing nothing more than … ’ might have slipped past the tough sub-editor more easily.
It only follows then, that AI is incapable of coming up with a slew of novel ideas and choosing only the most suitable ones to produce a work of art as a human would.
E46 makes a good choice at the start, and I think most readers will realise that this sentence is looking back at the whole passage on AI’s debit side. Personally, I might have been tempted to go even further and write ‘follows from these three disabilities,’ just to underline that point. I think ‘slew of novel ideas’ matches the literary flourish of the original, and ‘incapable of’ is a nice strong choice. E46 gets very close to showing that AI fails at every stage of the process here, but I could still imagine sceptical questions from some picky readers: ‘Do you mean AI can come up with a single idea sometimes? Do you mean it come up with some ideas, but just can’t chose between them? Do you mean that it has its own, non-human way of turning ideas into something?’. Most solutions I imagine here involve splitting this sentence into two with an ‘AI is incapable of ….. . It can neither ….nor …. nor …’ pattern.
In short, AI finds itself at a loss when it comes to producing original works. Image-generating AI can follow instructions to depict the Tokyo Tower in van Gogh’s style, and do just that. And yet, AI lacks the capacity to ever create a style all its own.
E46’s ‘at a loss’ really gets to the heart of what 難しい means for AI (cleverly installed as the subject of this sentence). The choice of ‘depict’ for 描いてis a great way to avoid problems over paint vs. draw. I liked ‘follow instructions,’ but I wondered if the final ‘do just that’ then becomes redundant. Also does ‘all its own’ draw a bit too much attention from ‘ever create a style’?
Of course, if you were to put a human to the same task, the reality is that they too would be hard-pressed to create a new style when working with age-old media such as paintings.
I think E46 does very well here, and I believe adding ‘same task’ is key to making this sentence work as well as it does. Just as a style point, I would have preferred ‘If people were put to the same task’. E46 has simplified the English version of this sentence, relative to the Japanese original, and that is often a good choice, but might we have lost the sense of an author-reader dialogue that (I felt) enlivened the original? I am not sure how E46 handled もはやhere; it looks as if it might have got dropped. However, I argue that the もはや concept is very important for the argument to proceed, showing that there was a time when we could create new styles of painting, even though we may not be able to do so now.
… the artistic possibilities of those media are pursued in all directions until they finally run dry.
E46 made a very effective etymological play in rendering ‘可能なアート’ as ‘artistic possibilities’ here. It hints at a future where we (mere, un-van-Gogh-like mortals) can do artistic things with new technologies, while preserving the link back to the actual art done by van Gogh with his oil paints.
Put another way, however, this means that as long as technology continues to advance, new genres of art will be born in turn, ensuring that opportunities to showcase humanity’s radical creativity will remain.
E46 has provided some of the drama I was hoping for here. He also managed to add a bit more literary flourish with ‘radical creativity,’ but it is slightly different from the way I would have interpreted the original message. I believe the point was more to do with the innovations that result when people are given opportunities for creative self-expression, rather than something we could describe as radical.
What is truly needed to excel in the age of AI is volition, experience, and judgment, as well as creativity that makes full use of technology. With schools training children to be test-taking robots and driving them to fill their minds with all the information they can handle, our current educational system is likely in need of drastic reform.
E46 handled the closing sentences well (but please see my general comments). I think it was a good idea to explore alternatives to ‘cram,’ and it works well to add ‘schools’. Stating ‘our education system’ nicely alerts readers to these being comments on education in the author’s home country. I believe E46’s best (innovative!) translation choice across these sentences was ‘excel,’ which gets close to the author’s vision for 活躍 in the economy of 2030.
E51
As with E46, E51 showed a great ability to match natural English expressions with the original meaning. I imagine that E51 too enjoys playing with language, based on her many imaginative choices. However, in a few places, the imaginative choices in her translation got a little too far from the original Japanese text. I should add there were other occasions where I felt E51 captured the author’s voice brilliantly.
AI and the Birth of the Idea-to-Product Economy
E51’s entry worked well as a title, and ‘birth’ could nicely leave the question of when this economy was/will be born open. As I said elsewhere, I think the title could be improved by finding words other than ‘idea’ and ‘product’.
The world is currently abuzz over ChatGPT, an AI text generator which can not only answer questions on any subject, but also generate essays, novels, and even computer programs.
For me, E51 produced by the far the best opening sentence. It is a very nice example of how to untangle the complexity around ‘not only …. but also’. To make this opening even stronger, how about ending the first sentence at ChatGPT, and starting a second sentence with ‘This AI text generator …’?
Of course, I cannot see into the author’s mind, but I have a hunch that ‘abuzz’ captures exactly how he sees the advent of ChatGPT. Having ‘The world’ as the subject prevented ChatGPT from dominating the opening on its own. The phrase ‘questions on any subject’ was an excellent choice for あらゆる質問.
E51 also generally does a good job in the second sentence (not reproduced here), especially with the choice of ‘meanwhile’ for 一方. As I commented elsewhere, I think ‘suitable’ needs a bit more work to establish it as the adjective specifically referring to the input-output relationship for image generation. I also feel ‘text prompt’ would be preferable to ‘verbal prompt’ for 言葉を入力する.
Using these sorts of “generative AI”, it's easy to create picture books and manga, and even sell the resulting works as e-books. Eventually, by connecting these systems to 3D printers, individual users may even be able to craft tools, accessories, and household items on the spot.
E51 has suddenly reverted to quote marks having correctly rejected them before. She has also changed her translation of 生成AI, when it would be preferable to keep the term consistent, especially in the early stages of the article. I do not think there is a big problem with the choice of ‘connecting’ for 組み合わせることbut it could hit at some automated, cabling system rather than a set-up where people are involved at every step. I believe ‘combining’ might work better. I have a similar reservation about ‘on the spot’ for たちどころに. Readers may imagine someone going from house to house to print out a sofa for each client, rather than the sort of ‘in-the-click-of-a-mouse’ type timing that I believe the author is actually thinking about here.
Conversely, as ordinary human skill becomes unable to keep pace with AI, it will become increasingly difficult to make a living as a professional artist.
I thought that E51 handled this very well; ‘unable to keep pace with’ and ‘difficult to make a living as a professional artist’ perfectly captured the complicated nuances here. I think it very cleverly introduced readers to artists but allows them to guess that this article will go on to discuss more than just the implications for artists.
Even today, the ratio of “Artist / Designer” job openings to applicants is only 0.18 – in other words, only around 1 artist in 5 is able to find work.
I liked the simplicity of ‘around 1 artist in 5’, but can those who do not find work really be described as artists? So sadly, I have to conclude that this phrase does not work. As a heads-up, I believe most journals that publish this sort of opinion piece would probably prefer numbers under ten to be spelled out. The quote marks could be misinterpreted if readers take them as indicators of irony (which is not impossible in this context), so I think we could polish up the way we describe this statistical category.
Ultimately, generative AI will impact not only artists, but also clerical workers, accountants, tax preparers, and even academic instructors and researchers such as myself. Virtually all white-collar professions will inevitably be affected.
E51 again showed a nice way to handle a ‘not only … but also’ structure naturally in English. Splitting into two sentences was also a good stylistic choice, but with two similar verbs (impact/affect), why not try something like ‘in the same boat’ for the second sentence. Turning 影響から逃れられない into ‘inevitably affected’ cannot be wrong, but I wonder if ‘must face up to the effects’ would have fitted the tenor of the article better, if the wording ‘cannot escape’ is to be rejected.
Current AI systems are lacking in three crucial elements: intention, experience, and judgment.
I have the same comment about the nice use of the colon, and the same question as to whether ‘judgement’ fully captures 価値判断, which I directed to E46.
I am not sure if ‘intention’ can work as an uncountable property here. I imagine using ‘have no intention’ only for a very specific action I have decided not to take. If we want to use ‘intention,’ then maybe we should consider converting this sentence to an action-describing pattern (e.g., AI cannot form intentions/gain experiences/make subjective judgments).
Generative AI such as ChatGPT could be described as ‘top-scoring elite job candidates’, but they are also passive and unmotivated employees. They lack the will and intention to begin work of their own volition or start new projects on their own.
The phrase ‘passive and unmotivated employees’ was an excellent choice here. The original phrase has real currency in the workplace, so this description fits perfectly. The term ‘top-scoring candidates’ is not too far off the mark either; I think deleting ‘elite’ would have been an improvement, and ‘entrance exam candidates’ might be more accurate than ‘job candidates’.
From this point on, E51 has switched to translating 生成AI as ‘generative AI’ and seems to treat it as a plural. I believe ‘AIs’ could be plural but not AI, but switching a key term like this halfway through a text (E51 also previously used ‘AI generator’ and ‘AI systems’) is the sort of thing that confuses readers.
Generative AI could also be compared to agoraphobic shut-ins, spending all their time browsing the web and reading books rather than ever leaving their homes.
This sentence reads very well in English, but I suspect ‘agoraphobic’ is going a little further than the original text justifies. Checking out the natural usage of 部屋にこもりきり in Japanese, it seems to be used in many situations (such as a student just before the exams) that don’t always correspond to cases of agoraphobia, as I understand a psychologist would diagnose it.
As with my previous point, most readers will surely see ‘AI’ as singular, so ‘their time’ might strike a discordant note (and continue to cause problems when readers come to ‘they’ in the next sentence).
Consequently, they are unable to create works as humans do, by brainstorming countless original ideas and picking out the ones we decide are actually desirable.
E51 gets this sentence off to a good start with ‘unable to create works as humans (aka ‘we’!) would,’ but the second half of the sentence has a bit of problem with ‘decide’; it is bit confusing whether ‘we’ or ‘they’ is the best subject for ‘decide’, which makes me think that neither approach is ideal. There is a strong case for splitting the sentence here into a pattern like this: ‘… AI models are unable …; they cannot ….’
I think E51 then handles the transition to Tokyo Tower fairly well, but I recommend her to think about my general comments at that point, too.
You may ask, isn't this difficult for human beings as well? Looking at the history of old media such as painting, we have no choice but to agree.
E51 makes an excellent choice in challenging readers with a rhetorical question here, and I think this perfectly captures the spirit of the original and preserves its liveliness (but should we have a colon, not a comma, after ‘ask’?). However, the second half of the sentence runs into a bit of trouble. It seems to suggest that if we look at the Old Masters’ paintings, we will find plenty of evidence of unoriginal people in bygone times. I know that is not what E51 has in mind, but it would seem to be the literal interpretation of the words on the page.
Paintbrushes and cameras are both technologies, and painting and photography are both forms of media.
E51 seems set on having a stand-off with the author about the definition of technology here! I am not sure I could define exactly what counts for technology in painting, and maybe E51 is even right. However, I suspect there is a reason that oil paints are mentioned here. Even a total non-artist such as myself knows that van Gogh developed an innovative style because he mixed recently discovered pigments into his oil paints (think about all those Cadmium Yellow sunflowers!), and so oil paints fit nicely into the sequence of the argument our author is building. If we really have concerns about the use of the word technology, perhaps we are better off with something like ‘oil pants represent an element of technology’ than seemingly trying to contradict the author.
New technologies create new media, and as we fumble around discovering what types of art we can create in a new medium, eventually we exhaust all of the possibilities.
I can see how E51 got to ‘fumble around’ from 模索, but I feel that this choice takes readers in the wrong direction. Based on a little background reading and my understanding of the text, I would say that people doing 模索 is the single most indispensable element in Professor Inoue’s economic vision, so our translation of this verb has to be capable of a positive nuance. The other finalists have gone for a mixture of ‘explore’ and ‘pursue,’ and I think those options work much better. However, I have to own up to having an unfair advantage here: my opinion here is partly based on reading the author’s latest book (AI失業：生成AIは私たちの仕事をどう奪うのか？), which was published one week after the deadline for submitting entries to this contest.
However, by the same token, as long as technology continues advancing, new artistic genres will continue to emerge, and humans will still have opportunities to demonstrate our innovation and creativity.
I have used underlining here to return to my point that ‘people’ or ‘we’ would be a more natural choice than ‘humans’ here, and make a better match for the excellent suggestion of ‘we’ in the previous sentence. However, I do not want to detract from this nice strong sentence, which maybe came closest to matching the force with which (I believe) the author wants to convey his message here. ‘By the same token’ is an imaginative choice for 逆に言うと, and the choice of active voice for the final clause is a good one.
In order to fully participate in the age of AI, we need intention, experience, and judgment; we need creativity that makes full use of the technology available to us.
Our current educational system, which pressures students to simply cram in knowledge and turns children into point-scoring machines, may be what is most in need of drastic reform.
E51’s final two sentences made a nice strong conclusion. The need for drastic reform has got a bit more convoluted than it needs to be (I guess this was the effect of だろう), but it does not really spoil the very nice job E51 did overall at the end. ‘Our’ helps to locate the education system in the author’s home country, too. Splitting the sentence on the age of AI into two powerful clauses separated by a semi-colon works well. I think readers will go away with that message ringing in their ears.
(Continued to Part 3 by Ruth McCreery)