Páxinas no tema: < [1 2 3 4 5 6 7] > | Who is not using MT today? Persoa que publicou o fío: Gary Evans
| Daryo United Kingdom Local time: 08:33 Serbian to English + ... You started believing in your own hype??? | Aug 16, 2019 |
"Not only does it prooduce frighteningly good results" you meant "[MT] produce frighteningly good results"? Yes, for sure - occasionally a bit more frequently than random typing would produce a meaningful sentence the only slight problem being that when you compare the nice sounding (/ "frighteningly good" as one would put it in a good marketing spiel) "translation" with the source text, more often than not you can find only tenu... See more "Not only does it prooduce frighteningly good results" you meant "[MT] produce frighteningly good results"? Yes, for sure - occasionally a bit more frequently than random typing would produce a meaningful sentence the only slight problem being that when you compare the nice sounding (/ "frighteningly good" as one would put it in a good marketing spiel) "translation" with the source text, more often than not you can find only tenuous connections between the two. That of course being a potential trap ONLY for the "nice sounding" parts, not for the MT output that is plain nonsense. Yes, sure - MT is "frighteningly good" - at randomly distorting the meaning of the ST, about as reliable as a witness who would carelessly and randomly throw a lie or two and from time to time inadvertently let slip the unadorned whole truth. But then, that's why this "frighteningly good" output of course ALWAYS needs to be checked? Or are we supposed to disregard that bit, as it's "frighteningly good"? As for "I also use it for teaching English" I'm sure a lot of language schools will want to get in touch with you, to learn first-hand how to dispense with language teachers ... Whatever this ISO standard has to say (I definitely won't be paying 118 Swiss Francs for the privilege of reading what had to say those who want to turn the art of translation into a mass production process) and however it is sensible to have another pair of eyes looking at your work, I know how it happens in practice. Just take a look at forum discussions about how useful "proofreaders/editors" are with human translators ... Gary Evans wrote: Hi Daryo, "...but there is a number of true professionals whose output is in no need of any "editing". May I just revise that fragment: "...but there are a number of true professionals whose output doesn't need any "editing". QED. I'm sure you're good, but ISO 17100 (previously EN 15038) requires revision (postediting if you like) as part of the translation process. You seem to think MT is only useful for agencies. It isn't, but you seem to be convinced by your own fallacy. Not only does it prooduce frighteningly good results, I also use it for teaching English. Regards, Gary Daryo wrote: Gary Evans wrote: Hi Fatih, Of course MT output needs to be post edited. Same goes with human translations, but MT is getting better and is much faster than us humans. And we're only just getting started! Nobody is considering dumping CAT tools anytime soon. In fact, CAT tools with integrated MT engines are now quite normal and commonly used across the industry. It's just that some translators are not admitting to using MT. It'S so useful, you'd be mad not to make use of it IMO. Expect MT to be doing the bulk of translations in the near future and plan for this, or you'll probably be looking for work elsewhere in the future. ... An output that "of course needs to be post edited" in my rulebook IS NOT a "translation" - at best it could be charitably called an " attempted translation" versus "Same goes with human translations" NO IT DOES NOT - it might come as a surprise to you (/ or maybe you have forgotten it) but there is a number of true professionals whose output is in no need of any "editing". Are you seriously trying to equate the two? MT being in essence a FAKE description of what is more accurately "machine produced text" that tries to be a translation, I can't see how anyone could "admit" to using it. With a bit of stretching, you could call "MT" any use of glossaries or dictionaries, or reusing fragments of your own previous work. The story of "Machine Translation" reminds me in some regards of the story of speed cameras - a technology initially created to help racing drivers improve their driving was perverted into a money making device for fleecing ordinary drivers. What is today called "MT" will at some point in future be good enough to produce an output that can honestly be called a "translation" but the way that unfinished product still not ready for release is used today by agencies, it is just an excuse to exploit unsuspecting translators by making them do translations at the rates of supposed "editing". ▲ Collapse | | | Reversed roles | Aug 16, 2019 |
Gary Evans wrote: BTW. I've discovered a neat solution in one respect. I sometimes use MT to automatically translate whole documents in one go and then sit with the client as we proofread them together. The client pays an hourly rate and benefits from lower costs, plus they get English lessons at the same time. Quick and dirty you may think, but it works really well as it offers a different business model for me. Bernhard Sulzer wrote: What are you trying to do here? Notwithstanding this strange setup where a client has the time and skill to work with you that way, why in the world would you charge less? It's still you, the human translator who judges the MT output and decides what's right or wrong, what is stylistically unacceptable, which sentence structures need to be fixed, which terms are completely wrong, etc. You are selling yourself cheap. providing an intellectual service that requires your "human" knowledge and skill" to make sure the translation is accurate. The machine is not going to do that for you. And that is essential, especially in the legal, medical and technical fields. I totally agree with Bernhard. Plus a question to Gary: If you are a freelance translator and you are living ONLY from providing translations full-time then why are you happy providing lower costs to your clients? Simply it doesn't make sense.![](https://cfcdn.proz.com/images/bb/smiles/icon_smile.gif) "benefits from lower costs" Lower costs = you have to work more to get the same amount of money as you used to -> so you'll be more tired -> so you'll make more mistakes -> and the quality of your work will decline. End of the story.
[Edited at 2019-08-16 07:52 GMT] | | | Tom Hoar (X) United States Local time: 03:33 English Observations of an MT technology vendor. | Aug 16, 2019 |
First, I'll to introduce myself. I'm the technology vendor who develops and sells Slate Desktop MT software. I'm not a translator (not even bilingual unless English-American counts). I've never worked for an LSP or agency. I don't come from academia with a PhD in machine translation. I came from an intelligence background where I supported translators with some unique (and very cool) ... See more First, I'll to introduce myself. I'm the technology vendor who develops and sells Slate Desktop MT software. I'm not a translator (not even bilingual unless English-American counts). I've never worked for an LSP or agency. I don't come from academia with a PhD in machine translation. I came from an intelligence background where I supported translators with some unique (and very cool) technologies. You can learn more about me here: Who I Am and Why Translation. I continued my language technology career with translators in the commercial translation services world. Now that my cards are on the table, I offer these observations about Gary's topic and all your comments. TONE: The tone of this conversation shows a significant shift from conversations only 2 years ago. Heck! even six months ago. In previous conversations, translator have attacked your colleagues for the mere mention of machine translation, much less the hint that they were using it. I'm not sure why the change, but to me, it's a welcomed. Christine's comments exemplify this shift (emphasis added). Christine Andersen wrote: I am not in principle opposed to MT any more. It is not going to go away, and it has its uses. Where MT comes from: Gary opened this topic saying: Gary Evans wrote: I've been testing machine translation tools for a while now... After reading all comments, it's clear that everyone uses the term "MT" in reference the online cloud-based services. From Gary's descriptions of his testings and everyone comments I'm pretty sure that you've all only use those online services. Until roughly 2008, "MT" for professionals to use was only available as desktop software from companies like SYSTRAN and PROMT. Then, along came newer technologies like the Internet, machine learning and artificial intelligence. Greed and desire for control moved machine translation technologies to online services. Desktop software applications nearly became extinct. It's my opinion that this migration to online services caused most of the problems that translators experience with MT. That's a subject for another topic. Confidentiality vs Privacy: Gary, you mentioned this about confidentiality. Gary Evans wrote: I appreciate concerns about confidentiality, but (naming no names) there are MT engines which explicitly do not store any data as part of their offer. The reality is, despite these online services' marketing literature, it's simply not true that they "do not store any data." When you read the fine print in their terms of use, the reality is very different. In addition, every online MT service must log every incoming source segment for their logs comply with law enforcement and counter-terrorism requirements. That said, confidentiality is not a professional translator's only concern (risk) when using online services. Online services can -- and do -- exploit those logs to make money, much like Facebook exploits your activity logs to make money. This exploitation has nothing to do with the segments you translate. The logs are a gold mine of your activities, preferences and habits. They expose your productivity (work over time). For example, the services can and do use the log to determine if you're working "too fast," which might be an indicator that you're creating low-quality translations. Many translators who use desktop CAT tools take the time to anonymize this information from packages before they submit their work to clients. You can't protect yourself with the online tools. John Fossey, thank you for pointing out the details in the Wikipedia Postediting article. I read the underlying PDF files and just now edited the article to more accurately relay the research content (emphasis added here for clarity): What Wikipedia said (before my edits): ... some academic studies suggest that time savings under realistic working conditions are more likely to be between 15–40%. Several professionals claim that it takes longer to make correction than to translate from scratch. What Wikipedia now says: ... some academic studies suggest that time savings under actual working conditions are more likely to be between 0–20%. Professionals have also reported negative productivity gains where corrections require more time than to translate from scratch. In summary, much of what translators know about MT comes from companies that stand to gain (secure a continuous income stream of subscriptions) when you and/or your clients use their services. I call these companies the MT mafia (alternate mafia definition: a closed group of people in a particular field, having a controlling influence). Regardless of what you think of MT as a technology, it's worthwhile to dig deeper and see how you and others profit (or not) from your use of their services. ▲ Collapse | | | Gary Evans Germany Local time: 09:33 German to English INICIO DE TEMA MT is here to stay for good, or bad (for some). | Aug 16, 2019 |
Hi Tom, Thanks for your thoughts. I just checked the T&C of my favourite MT and it does claim to protect the data. The EU has introduced tough legislation about data protection, unlike in the USA. Maybe that helps. I use MT for translations as well as in my teaching work. Here's what one translator posted on this thread as a response: I'm sure a lot of language schools will want to get in touch with you, to learn first-hand how to dispense with languag ... See more Hi Tom, Thanks for your thoughts. I just checked the T&C of my favourite MT and it does claim to protect the data. The EU has introduced tough legislation about data protection, unlike in the USA. Maybe that helps. I use MT for translations as well as in my teaching work. Here's what one translator posted on this thread as a response: I'm sure a lot of language schools will want to get in touch with you, to learn first-hand how to dispense with language teachers ... This demonstrates that some people just can't debate a topic without sinking to ad hom attacks. No, AI isn't going to take away your job just yet, but the world is changing rapidly now. IMO students benefit from translating. I've been doing this for some years now. One fun thing to do is to use MT to translate their English essays etc. into German. I then translate this output back into English using the MT and compare and contrast. Many classic errors are corrected in the TM. I can then use my CAT tool to write comments and provide links to lessons on those classic errors. It's fun, even though the MT was not designed for this purpose. I'd like to develop a way of getting AI to label these error categories. This is an open source project on Gitlab BTW. I don't think I'll be out of teaching work anytime soon I hope. Who knows with AI though. Doctors, lawyers and other previously safe professions are changing rapidly as AI matures. I can only hope that the human touch will be needed as AI gets better and medical diagnosis, legal advice and helping people learn. Kind regards, Gary Tom Hoar wrote: First, I'll to introduce myself. I'm the technology vendor who develops and sells Slate Desktop MT software. I'm not a translator (not even bilingual unless English-American counts). I've never worked for an LSP or agency. I don't come from academia with a PhD in machine translation. I came from an intelligence background where I supported translators with some unique (and very cool) technologies. You can learn more about me here: Who I Am and Why Translation. I continued my language technology career with translators in the commercial translation services world. Now that my cards are on the table, I offer these observations about Gary's topic and all your comments. TONE: The tone of this conversation shows a significant shift from conversations only 2 years ago. Heck! even six months ago. In previous conversations, translator have attacked your colleagues for the mere mention of machine translation, much less the hint that they were using it. I'm not sure why the change, but to me, it's a welcomed. Christine's comments exemplify this shift (emphasis added). Christine Andersen wrote: I am not in principle opposed to MT any more. It is not going to go away, and it has its uses. Where MT comes from: Gary opened this topic saying: Gary Evans wrote: I've been testing machine translation tools for a while now... After reading all comments, it's clear that everyone uses the term "MT" in reference the online cloud-based services. From Gary's descriptions of his testings and everyone comments I'm pretty sure that you've all only use those online services. Until roughly 2008, "MT" for professionals to use was only available as desktop software from companies like SYSTRAN and PROMT. Then, along came newer technologies like the Internet, machine learning and artificial intelligence. Greed and desire for control moved machine translation technologies to online services. Desktop software applications nearly became extinct. It's my opinion that this migration to online services caused most of the problems that translators experience with MT. That's a subject for another topic. Confidentiality vs Privacy: Gary, you mentioned this about confidentiality. Gary Evans wrote: I appreciate concerns about confidentiality, but (naming no names) there are MT engines which explicitly do not store any data as part of their offer. The reality is, despite these online services' marketing literature, it's simply not true that they "do not store any data." When you read the fine print in their terms of use, the reality is very different. In addition, every online MT service must log every incoming source segment for their logs comply with law enforcement and counter-terrorism requirements. That said, confidentiality is not a professional translator's only concern (risk) when using online services. Online services can -- and do -- exploit those logs to make money, much like Facebook exploits your activity logs to make money. This exploitation has nothing to do with the segments you translate. The logs are a gold mine of your activities, preferences and habits. They expose your productivity (work over time). For example, the services can and do use the log to determine if you're working "too fast," which might be an indicator that you're creating low-quality translations. Many translators who use desktop CAT tools take the time to anonymize this information from packages before they submit their work to clients. You can't protect yourself with the online tools. John Fossey, thank you for pointing out the details in the Wikipedia Postediting article. I read the underlying PDF files and just now edited the article to more accurately relay the research content (emphasis added here for clarity): What Wikipedia said (before my edits): ... some academic studies suggest that time savings under realistic working conditions are more likely to be between 15–40%. Several professionals claim that it takes longer to make correction than to translate from scratch. What Wikipedia now says: ... some academic studies suggest that time savings under actual working conditions are more likely to be between 0–20%. Professionals have also reported negative productivity gains where corrections require more time than to translate from scratch. In summary, much of what translators know about MT comes from companies that stand to gain (secure a continuous income stream of subscriptions) when you and/or your clients use their services. I call these companies the MT mafia (alternate mafia definition: a closed group of people in a particular field, having a controlling influence). Regardless of what you think of MT as a technology, it's worthwhile to dig deeper and see how you and others profit (or not) from your use of their services.
▲ Collapse | |
|
|
Gary Evans Germany Local time: 09:33 German to English INICIO DE TEMA Money doesn't make the world go round | Aug 16, 2019 |
Hi Katalin, A friend asked me to translate a paper he had written for publishing in a journal. It would have cost him at least € 500 which he didn't have. I translated it using an MT engine and we then sat together and postedited it. The topic, climate change denial, was fascinating. It's just been through peer review and published. I did it pro bono (okay received a bottle of wine for my services). Another acquaintance needed a document translated quickly for a conference. She's ... See more Hi Katalin, A friend asked me to translate a paper he had written for publishing in a journal. It would have cost him at least € 500 which he didn't have. I translated it using an MT engine and we then sat together and postedited it. The topic, climate change denial, was fascinating. It's just been through peer review and published. I did it pro bono (okay received a bottle of wine for my services). Another acquaintance needed a document translated quickly for a conference. She's working on the UN'S 17 Sustainable Development Goals. We both learned a lot while postediting this and I got to earn €100 for an hour's work. The coffee was free btw. But I don't just work as a translator. I also teach and am an activist in my free time. Regards, Gary Katalin Szilárd wrote: Gary Evans wrote: BTW. I've discovered a neat solution in one respect. I sometimes use MT to automatically translate whole documents in one go and then sit with the client as we proofread them together. The client pays an hourly rate and benefits from lower costs, plus they get English lessons at the same time. Quick and dirty you may think, but it works really well as it offers a different business model for me. Bernhard Sulzer wrote: What are you trying to do here? Notwithstanding this strange setup where a client has the time and skill to work with you that way, why in the world would you charge less? It's still you, the human translator who judges the MT output and decides what's right or wrong, what is stylistically unacceptable, which sentence structures need to be fixed, which terms are completely wrong, etc. You are selling yourself cheap. providing an intellectual service that requires your "human" knowledge and skill" to make sure the translation is accurate. The machine is not going to do that for you. And that is essential, especially in the legal, medical and technical fields. I totally agree with Bernhard. Plus a question to Gary: If you are a freelance translator and you are living ONLY from providing translations full-time then why are you happy providing lower costs to your clients? Simply it doesn't make sense. "benefits from lower costs" Lower costs = you have to work more to get the same amount of money as you used to -> so you'll be more tired -> so you'll make more mistakes -> and the quality of your work will decline. End of the story. [Edited at 2019-08-16 07:52 GMT] ▲ Collapse | | | We are here to make business: translation is our profession. | Aug 16, 2019 |
Gary Evans wrote: Hi Katalin, A friend asked me to translate a paper he had written for publishing in a journal. It would have cost him at least € 500 which he didn't have. I translated it using an MT engine and we then sat together and postedited it. The topic, climate change denial, was fascinating. It's just been through peer review and published. I did it pro bono (okay received a bottle of wine for my services). Another acquaintance needed a document translated quickly for a conference. She's working on the UN'S 17 Sustainable Development Goals. We both learned a lot while postediting this and I got to earn €100 for an hour's work. The coffee was free btw. But I don't just work as a translator. I also teach and am an activist in my free time. Regards, Gary Katalin Szilárd wrote: Gary Evans wrote: BTW. I've discovered a neat solution in one respect. I sometimes use MT to automatically translate whole documents in one go and then sit with the client as we proofread them together. The client pays an hourly rate and benefits from lower costs, plus they get English lessons at the same time. Quick and dirty you may think, but it works really well as it offers a different business model for me. Bernhard Sulzer wrote: What are you trying to do here? Notwithstanding this strange setup where a client has the time and skill to work with you that way, why in the world would you charge less? It's still you, the human translator who judges the MT output and decides what's right or wrong, what is stylistically unacceptable, which sentence structures need to be fixed, which terms are completely wrong, etc. You are selling yourself cheap. providing an intellectual service that requires your "human" knowledge and skill" to make sure the translation is accurate. The machine is not going to do that for you. And that is essential, especially in the legal, medical and technical fields. I totally agree with Bernhard. Plus a question to Gary: If you are a freelance translator and you are living ONLY from providing translations full-time then why are you happy providing lower costs to your clients? Simply it doesn't make sense. "benefits from lower costs" Lower costs = you have to work more to get the same amount of money as you used to -> so you'll be more tired -> so you'll make more mistakes -> and the quality of your work will decline. End of the story. [Edited at 2019-08-16 07:52 GMT] Hi Gary, You did the right thing concerning the pro bono job: if somebody wants to translate for free to help somebody or others, she/he can do so in his/her spare time on her/his own. The less members are involved, the more sure that your job will be received by those who really need them without any misuse. Concerning money: you forget something very important here. Many professional translators are here on this site to make money because that's what we are expert in: translation, specialized translation (medical, legal, business, technical etc.) where human logic is essential. This is our profession, our career, our business (!), our mission. Many of us love to create new terms, new words, able to see the big picture of a text. Once I wrote in a post and I still think: sooner or later those people who are doing MT post-editing will lose their ability to translate as a human and to think logically. Their job will be monotone and dull. Only a true linguist will understand what I'm telling, because a true linguist's job is also her/his hobby. + Added: There is a huge difference between 1) making a text comprehensible vs. 2) creating a fluent and extremely accurate text that doesn't smell like a translation and if it's read by other experts (physicians, lawyers and engineers etc.) they won't be able to tell that it was a translation. + Added: https://www.proz.com/forum/machine_translation_mt/318856-rates_per_hour_for_mtpe.html#2704064 I would never make a post-editing. I love translate, to create, to explore the text by myself, I'm passionate about translation. That's why I don't do proofreading either, except if I know the translator very well. Bests, Katalin
[Edited at 2019-08-16 16:21 GMT]
[Edited at 2019-08-16 16:38 GMT] | | | Tom Hoar (X) United States Local time: 03:33 English The devil's in the details | Aug 16, 2019 |
Gary Evans wrote: I just checked the T&C of my favourite MT and it does claim to protect the data. The devil's in the details and lawyers write the details. I've never read a T&C without clauses that give the service provider an "out" that allows them to share the data with whomever they choose. Since you're working with German, let's look at this paragraph from DeepL's T&C (emphasis added). DeepL T&C: 3.6 DeepL will only store Content on its servers to the extent technically required to provide its services. Customer’s Content will neither be perpetually stored on DeepL’s servers nor returned to the Customer. For the avoidance of doubt, DeepL shall be entitled to create and retain access logs for billing, security and statistical purposes. Such access logs shall not contain any of Customer’s Content. However, access logs may contain meta data of API Requests such as time of the API Request and size of the transmitted Content. - To the extent: who determines that extent? DeepL
- perpetuity: ok, so a 100-year is not in perpetuity.
- shall be entitled: speaks for itself.
This is the "out" that DeepL can use to share your data with whomever they like as long as they claim it's necessary to provide their services. DeepL T&C: 6.4 All rights regarding Customer’s Content remain with Customer. However, Customer grants DeepL the non-exclusive worldwide right to use the Content solely in order to provide DeepL’s services to the Customer. In particular, Customer grants to DeepL the right to temporarily store, modify, process, translate and transmit the Content, and to sublicense the foregoing rights to its subcontractors, to the extent required to provide the services set out in this Agreement. Regarding your T&C that promises to protect the data, I ask you "protect from whom?" It certainly does not prevent the company from using the data for its own benefit. Yes, I'm familiar with the GDPR. Top 10 GDPR Violations and Incidents of 2018. | | | Dan Lucas United Kingdom Local time: 08:33 Membro (2014) Japanese to English For sensitive documents, just say no | Aug 18, 2019 |
Tom Hoar wrote: It certainly does not prevent the company from using the data for its own benefit. And this is basically the conclusion that I believe my clients have collectively drawn. Of course, the MT vendors argue that there's no problem, but it's a very murky area. The only way to be sure, at this stage, is not to use MT for sensitive documents. On twitter an academic recently asserted that you'd have to be "desperate" to ban MT. No, companies don't have to be "desperate" to forbid the use of MT, just keenly aware of the potential repercussions. If the use of MT were to lead to the unplanned disclosure of a company's material nonpublic information (or even the appearance of the same), the regulators would be kicking down the door of the company itself, not the doors of the MT vendors or of starry-eyed academics who've never worked in industry, or even of touchingly naive freelancers. The risk of such problems may be small, but the potential damage could be huge, all for the sake of saving a few tens or hundreds of dollars on an important document. Regards, Dan | |
|
|
Gary Evans Germany Local time: 09:33 German to English INICIO DE TEMA Then don't get lost with the devil | Aug 18, 2019 |
Hi Tom, So your argument against MT comes down to data security, but not its value as a CAT tool. Of course sensitive data is best kept safe, but a lot of the stuff I translate is for public consumption. For example, the CAT tool I have developed for my teaching work was designed for translating open source material i.e. Wikipedia articles etc. You should read my T&Cs![](https://cfcdn.proz.com/images/bb/smiles/icon_smile.gif) Regards, Gary Tom Hoar wrote: Gary Evans wrote: I just checked the T&C of my favourite MT and it does claim to protect the data. The devil's in the details and lawyers write the details. I've never read a T&C without clauses that give the service provider an "out" that allows them to share the data with whomever they choose. Since you're working with German, let's look at this paragraph from DeepL's T&C (emphasis added). DeepL T&C: 3.6 DeepL will only store Content on its servers to the extent technically required to provide its services. Customer’s Content will neither be perpetually stored on DeepL’s servers nor returned to the Customer. For the avoidance of doubt, DeepL shall be entitled to create and retain access logs for billing, security and statistical purposes. Such access logs shall not contain any of Customer’s Content. However, access logs may contain meta data of API Requests such as time of the API Request and size of the transmitted Content. - To the extent: who determines that extent? DeepL
- perpetuity: ok, so a 100-year is not in perpetuity.
- shall be entitled: speaks for itself.
This is the "out" that DeepL can use to share your data with whomever they like as long as they claim it's necessary to provide their services. DeepL T&C: 6.4 All rights regarding Customer’s Content remain with Customer. However, Customer grants DeepL the non-exclusive worldwide right to use the Content solely in order to provide DeepL’s services to the Customer. In particular, Customer grants to DeepL the right to temporarily store, modify, process, translate and transmit the Content, and to sublicense the foregoing rights to its subcontractors, to the extent required to provide the services set out in this Agreement. Regarding your T&C that promises to protect the data, I ask you "protect from whom?" It certainly does not prevent the company from using the data for its own benefit. Yes, I'm familiar with the GDPR. Top 10 GDPR Violations and Incidents of 2018. | | | Gary Evans Germany Local time: 09:33 German to English INICIO DE TEMA Touchingly naive freelancers... | Aug 18, 2019 |
Hi Dan, Condescending remarks don't really help anyone here. Increasing numbers of people are making use of MT. The genie is out of the bottle. I've found MT to be useful in my teaching work as well as when translating. It offers new business models as well providing the world with a great tool to improve how we communicate. Businesses are using it to provide instant translations of customer reviews. It works! Concerns about data security are insufficient to prevent the rise o... See more Hi Dan, Condescending remarks don't really help anyone here. Increasing numbers of people are making use of MT. The genie is out of the bottle. I've found MT to be useful in my teaching work as well as when translating. It offers new business models as well providing the world with a great tool to improve how we communicate. Businesses are using it to provide instant translations of customer reviews. It works! Concerns about data security are insufficient to prevent the rise of MT. Translators should be aware of this fact and act accordingly. Regards, Gary Dan Lucas wrote: Tom Hoar wrote: It certainly does not prevent the company from using the data for its own benefit. And this is basically the conclusion that I believe my clients have collectively drawn. Of course, the MT vendors argue that there's no problem, but it's a very murky area. The only way to be sure, at this stage, is not to use MT for sensitive documents. On twitter an academic recently asserted that you'd have to be "desperate" to ban MT. No, companies don't have to be "desperate" to forbid the use of MT, just keenly aware of the potential repercussions. If the use of MT were to lead to the unplanned disclosure of a company's material nonpublic information (or even the appearance of the same), the regulators would be kicking down the door of the company itself, not the doors of the MT vendors or of starry-eyed academics who've never worked in industry, or even of touchingly naive freelancers. The risk of such problems may be small, but the potential damage could be huge, all for the sake of saving a few tens or hundreds of dollars on an important document. Regards, Dan ▲ Collapse | | | Gary Evans Germany Local time: 09:33 German to English INICIO DE TEMA Will AI take away your job? | Aug 18, 2019 |
[quote]Daryo wrote: "As for "I also use it for teaching English" I'm sure a lot of language schools will want to get in touch with you, to learn first-hand how to dispense with language teachers ... [quote] Hi Daryo, This is a typically defensive negative response to the suggestion that AI can help improve how we (translators, teachers etc.) work. I don't see AI threatening language teachers anytime soon. In fact, AI offers vast potential for improving both teaching and learning. The university where I teach is extremely interested in the potential of such tools. So how can AI help? MT can be used in various teaching scenarios. It can provide feedback to a language learner's output (then compare and contrast in groups). It can be used to improve written texts by translating from L1 to L2 to L1 for example. Natural language Processing (NLP), another emerging AI tool with great potential, can help students to improve their speaking skills. All of this is supplemented by the input of both teachers and learners. Feedback from my students has convinced me that this approach is both innovative and productive. That's how I see us benefiting from AI. Of course there will be those who will never use it. That's fine and why not? But others are moving in new directions as AI offers the potential for us to work with it for our benefit. Of course, AI can also be used to only benefit corporations, but as with any disruptive technology, we must ensure it works for us instead. Regards, Gary BTW if you find any other typos here, you're welcome to correct me! | | | Different professions, different working materials, different skills | Aug 18, 2019 |
[quote]Gary Evans wrote: [quote]Daryo wrote: "As for "I also use it for teaching English" I'm sure a lot of language schools will want to get in touch with you, to learn first-hand how to dispense with language teachers ... Hi Daryo, This is a typically defensive negative response to the suggestion that AI can help improve how we (translators, teachers etc.) work. I don't see AI threatening language teachers anytime soon. In fact, AI offers vast potential for improving both teaching and learning. The university where I teach is extremely interested in the potential of such tools. So how can AI help? MT can be used in various teaching scenarios. It can provide feedback to a language learner's output (then compare and contrast in groups). It can be used to improve written texts by translating from L1 to L2 to L1 for example. Natural language Processing (NLP), another emerging AI tool with great potential, can help students to improve their speaking skills. All of this is supplemented by the input of both teachers and learners. Feedback from my students has convinced me that this approach is both innovative and productive. That's how I see us benefiting from AI. Of course there will be those who will never use it. That's fine and why not? But others are moving in new directions as AI offers the potential for us to work with it for our benefit. Of course, AI can also be used to only benefit corporations, but as with any disruptive technology, we must ensure it works for us instead. Regards, Gary BTW if you find any other typos here, you're welcome to correct me!
Gary, I'm sorry to say but I'm not interested in teaching and I don't think I'm the only one here. This is a translation related site for translators. Many of us are professional translators, many of us are not translating Wikipedia and other public sites and many of us are not teaching. You also mentioned NLP and AI and helping students. We are translating, not teaching. This is our profession: translating highly medical, legal, business, technical, etc. texts. Actually I don't understand this hype now and convincing others about it. If somebody is happy using these tools then it's OK and I think it's his/her own business. Bests, Katalin
[Edited at 2019-08-18 10:38 GMT] | |
|
|
Gary Evans Germany Local time: 09:33 German to English INICIO DE TEMA AI is coming your way, whether translator, or teacher | Aug 18, 2019 |
Hi Katalin, This really is my point. AI is here to stay and it's changing the way we live, learn and work. It's not hype. Just look at the way some here are spooked by MT! Why is that? If MT is effective then it will be used increasingly. If it's all hype, then it won't affect the translation world. Regards, Gary | | |
Gary Evans wrote: Hi Katalin, This really is my point. AI is here to stay and it's changing the way we live, learn and work. It's not hype. Just look at the way some here are spooked by MT! Why is that? If MT is effective then it will be used increasingly. If it's all hype, then it won't affect the translation world. Regards, Gary Gary, It is a hype. This is a programmed hype. Why are you trying to convince people about that these tools are so good? If I found a good tool, I'd tell it to a few people, or give all the information in a thread to everybody, but I wouldn't use my spare time to convince others about using them... actually it would be an advertising. Bests, Katalin
[Edited at 2019-08-18 16:00 GMT] | | | Michele Fauble United States Local time: 00:33 Membro (2006) Norwegian to English + ... “AI is very, very stupid” | Aug 18, 2019 |
| | | Páxinas no tema: < [1 2 3 4 5 6 7] > | Non hai ningún moderador asignado específicamente a este foro. Para informar do incumprimento das normas do sitio ou para obter axuda, por favor, contacte co persoal do sitio » Who is not using MT today? Protemos translation business management system | Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!
The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.
More info » |
| TM-Town | Manage your TMs and Terms ... and boost your translation business
Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |