| Pages sur ce sujet: < [1 2 3 4 5 6] > | About the constant AI training job posts. Auteur du fil: Javi Tazinafo
|
|---|
Javi Tazinafo wrote:
This conversation is about AI training job posts flooding this platform. My point is they shouldn't be allowed here.
Especially when there are many AI freelancing jobs platforms out there, like e2f, DataAnnotation, Datavio, Neevo, Telus, and more.
I have done gigs on a few of those IA freelancing platforms, as I already know what they're about.
On those freelancing platforms, there aren't any translation/interpreting/subtitling jobs posted. You'll find jobs related to text editing/rewriting for AI, as well as sentiment analysis and query evaluations—all jobs related purely with AI whether it's text, images, or speech recording.
So, why those jobs that aren't purely related to translation/interpreting/subtitling are in here?
The freelancers interested in those jobs can go to those platforms, as I did. I don't see validation to allow those job posts in here.
[Edited at 2024-10-14 19:27 GMT]
| | | | David Mossop 
Royaume-Uni
Local time: 23:09
Membre (2010)
allemand vers anglais
| Not really that baffling | Oct 15, 2024 |
Baran Keki wrote:
Miranda Drew wrote:
In my language pairs (IT-EN), many end clients are directly using AI and completely cutting out translators and agencies. It is not just a "tool" (which I won't use unless the job specifies it. Correcting AI or MT translation takes me MORE time than just translating).
I know many agencies that have gone out of business, and tons of translators who've moved on to other fields.
One IT-EN translator I know in these forums has been so busy with work since the beginning of this year that he stopped posting here and some people actually thought he died in his flat in London.
Or it could be that AI replaced him, and he got turned into some kind of robotic being and lost his human qualities.
I appreciate that you are being tongue-in-cheek here, but your anecdote doesn't in any way discredit Miranda's observations, which completely chime with my own experience in my language pair (DE-EN). It's far from baffling - this is happening to many, many experienced translators. Most of the agencies I have worked with over the past 15 years - all of them reputable boutiques that have always valued and fairly remunerated high-quality work - have either gone bust or made it clear to me that human translation is now a thing of the past as far as they are concerned, and that MTPE is all they can now offer me.
So kudos and continued good fortune to your acquaintance and anyone else still making a decent living exclusively through human translation - long may it continue. But I don't think their individual success can be regarded as evidence for the fact that the industry as we knew it isn't in the process of being completely swept away.
| | | | | into ENG is feeling it | Oct 15, 2024 |
David Mossop wrote:
Miranda Drew wrote:
In my language pairs (IT-EN), many end clients are directly using AI and completely cutting out translators and agencies.
I know many agencies that have gone out of business, and tons of translators who've moved on to other fields.
...chime with my own experience in my language pair (DE-EN). It's far from baffling - this is happening to many, many experienced translators. Most of the agencies I have worked with over the past 15 years ..have either gone bust or made it clear to me that human translation is now a thing of the past
Ditto FR-EN from where I'm standing. "Into English" is definitely feeling the effects. Two causes I suspect.
1. Lots and lots of AI training content available in English. Far more so than, say, Dutch.
2. As Eng is something of a lingua franca, it's quite often used as a tool to communicate between people who don't speak it as a first language. And it has many varieties & flavours. Result, a lot of pretty ropey Eng is deemed acceptable by people who don't know any better. Far more than, say, Dutch.
In terms of banning or not banning anything, anywhere, I think 2 factors are important:
a) identifying clearly and precisely what is, and is not, covered by the ban
b) enforcing it.
I suspect a) will be the tricky bit. As has been mentioned, job posts will often bait and switch, as it were.
| | | | | Identifying those job posts | Oct 15, 2024 |
Charlie Bavington wrote:
In terms of banning or not banning anything, anywhere, I think 2 factors are important:
a) identifying clearly and precisely what is, and is not, covered by the ban
b) enforcing it.
I suspect a) will be the tricky bit. As has been mentioned, job posts will often bait and switch, as it were.
I think the way to identify a job post that isn't purely related to translation/interpretation/subtitling is by reading the whole post. So, perhaps it would take time, but it would be worth it, IMHO.
Sometimes, it might not even be necessary to read the whole post since titles are clear, like these on the job board right now:
* “Afrasia| Data Collection - Multiple Languages” > Job post content says: “Required service: Data Collection.”
* “Obsidian Image Understanding RAI English Demonstr” > Job post content says: “This project aims to evaluate whether the provided images, Human prompts, Agent responses, and a combination of the three violate the provided category. Service: Annotation”
* “Internet Ads Assessor | Portuguese Speakers in Brazil (WFH)” > Job post content says: “In this role, you will be reviewing online advertisements by rating them on their relevance to the search terms used as well as providing feedback on their language and cultural relevance in order to improve their content, quality, and layout”
None are looking for translators/interpreters/captioners because they aren't translation/interpreting/subtitling jobs. They look for multilingual freelancers. Again, I don't see why these job posts are allowed here.
| | |
|
|
|
Javi Tazinafo
Brésil
Local time: 20:09
Membre (2010)
anglais vers portugais
AUTEUR DU FIL | Naming the "fair" rate is about replacing "market value" with a more realistic concept | Oct 15, 2024 |
ibz wrote:
Setting a minimum price seems difficult to me as a word price that is ok for someone from a country where wages and livings costs are quite low is not acceptable at all for someone from a high-price country. How do you want to solve this problem?
This rate would be just so the job got the fair market seal. I don't think companies should feel obligated to fulfil those requirements, only if they want to distinguish themselves. I think defining that rate should be a process, and it should start with one that made a decent living across all poor and developing countries. My idea, as I said before, is USD 0.06/w. If an agency genuinely can't afford paying this much, it's because they themselves are being exploited, therefore they can't get the seal (which is fair). An agency that can pay this much but doesn't is arguably just profiting too much, therefore won't get the seal. I named this rate simply by disregarding what the market "wants" to pay and considering what would make for a decent living in most countries. It's a blanket rate by design.
| | | | Javi Tazinafo
Brésil
Local time: 20:09
Membre (2010)
anglais vers portugais
AUTEUR DU FIL | Bait and switch | Oct 15, 2024 |
Charlie Bavington wrote:
I suspect a) will be the tricky bit. As has been mentioned, job posts will often bait and switch, as it were.
That is exactly why I've been moving away from the idea of banning throughout the last few days. If you ban, they will post fake jobs to bait and switch later. If you don't ban and instead leave the chaos and exploitation for those who want it and open a lane for the distinguished ones, we could have a pocket of likeminded translators and agencies amidst the chaos, where we could get decent jobs. And the sweatshop agencies won't be inclined to lie into the fair market lane because they will find people who tolerate their crap outside of our lane.
| | | | | It was not censorship, please start a forum with this topic | Oct 16, 2024 |
Lieven Malaise wrote:
Javi Tazinafo wrote:
I don't think you did anything wrong, Monica.
Indeed, she didn't. She just expressed her opinion and me and a few others reacted to it. That's what these forums should be all about instead of nagging about on- or off-topic. I didn't agree with her opinion, but it was interesting to discuss it, so it's a shame that it has been censored.
It was not censorship, it is ensuring we stay on topic, which is not only one of the rules of the forums, but in the best interest of the member who started the topic. You are free to start a new forum to discuss the topic you feel is being censored. I'll not "censor" a post in this forum inviting the participants here to join your forum.
| | | | Lieven Malaise
Belgique
Local time: 00:09
Membre (2020)
français vers néerlandais
+ ...
Monica Oliveira wrote:
It was not censorship, it is ensuring we stay on topic, which is not only one of the rules of the forums, but in the best interest of the member who started the topic. You are free to start a new forum to discuss the topic you feel is being censored. I'll not "censor" a post in this forum inviting the participants here to join your forum.
Sometimes a thread goes in different directions. That's why forum members sometimes talk about one thing, to continue in a following paragraph starting with "on-topic:'. It happens all the time and it keeps things interesting, imo. It's hard to refrain from giving your opinion (for the good or for the bad) about AI in any AI topic, so why wouldn't you just allow it? It still doesn't make sense to me.
I appreciate that you are doing your job as a moderator, but if this is really the moderation standard, you must be the only one who is living by it that strictly.
| | |
|
|
|
| If note banned at least clearly marked | Oct 16, 2024 |
Javi Tazinafo wrote:
Charlie Bavington wrote:
I suspect a) will be the tricky bit. As has been mentioned, job posts will often bait and switch, as it were.
That is exactly why I've been moving away from the idea of banning throughout the last few days. If you ban, they will post fake jobs to bait and switch later. If you don't ban and instead leave the chaos and exploitation for those who want it and open a lane for the distinguished ones, we could have a pocket of likeminded translators and agencies amidst the chaos, where we could get decent jobs. And the sweatshop agencies won't be inclined to lie into the fair market lane because they will find people who tolerate their crap outside of our lane.
Then, if banning isn't a good option, I think these job posts should have a clear mark indicating they aren't purely about translation/interpreting/subtitling and thus intended for any freelancer with language knowledge.
They should have an exclamation mark to draw attention to the fact that these jobs aren't exclusively for professional translators/interpreters/captioners.
That way will be clear for professionals and also will show how many of them are posted. So we'll all have a clear image of the impact that AI has on the linguistic field and on downgrading the translation field by not employing professionals.
| | | | Javi Tazinafo
Brésil
Local time: 20:09
Membre (2010)
anglais vers portugais
AUTEUR DU FIL | Speaking of which | Oct 16, 2024 |
Lieven Malaise wrote:
Monica Oliveira wrote:
It was not censorship, it is ensuring we stay on topic, which is not only one of the rules of the forums, but in the best interest of the member who started the topic. You are free to start a new forum to discuss the topic you feel is being censored. I'll not "censor" a post in this forum inviting the participants here to join your forum.
Sometimes a thread goes in different directions. That's why forum members sometimes talk about one thing, to continue in a following paragraph starting with "on-topic:'. It happens all the time and it keeps things interesting, imo. It's hard to refrain from giving your opinion (for the good or for the bad) about AI in any AI topic, so why wouldn't you just allow it? It still doesn't make sense to me.
I appreciate that you are doing your job as a moderator, but if this is really the moderation standard, you must be the only one who is living by it that strictly.
This moderation discussion is also not on topic. Let's get back to it, please.
| | | | | More on identification | Oct 16, 2024 |
Maria Laura Curzi wrote:
Charlie Bavington wrote:
In terms of banning or not banning anything, anywhere, I think 2 factors are important:
a) identifying clearly and precisely what is, and is not, covered by the ban
b) enforcing it.
I suspect a) will be the tricky bit. As has been mentioned, job posts will often bait and switch, as it were.
I think the way to identify a job post that isn't purely related to translation/interpretation/subtitling is by reading the whole post. So, perhaps it would take time, [ examples snipped]
None are looking for translators/interpreters/captioners because they aren't translation/interpreting/subtitling jobs.
All fair points, but not quite what I meant. By bait-and-switch, I mean the kind of ad which would say "Translation job: blah" but then, when you apply, they reply "Sorry, that job has gone, but we have this opportunity... [e.g. your advert assessor / image understanding examples]".
Adverts which I suspect we would see more often if this site attempted to enforce any rules against the transparent version of the adverts we see now. Beware unintended consequences, and all that.
| | | | | Some considerations to help the discussion | Oct 16, 2024 |
Javi Tazinafo wrote:
I don't think you did anything wrong, Monica.
Guys, let's focus. We have site staff joining our conversation to take in ideas, offer points of view and bring our concerns to her team. That's the best we could have hoped for for this thread. Let's show them that we are not just an angry mob, that we can be pleased and that we can be listened to. There are a lot of unreasonable people out there. Let's show ProZ that they would rather deal with us then with them. Someone dropped the "So You're Saying" bomb on this thread and derailed the conversation. Nothing productive has ever happened after a SYS bomb, so let's not do that kind of thing again. Eye on the ball.
Let me know your thoughts. Eye on the ball, PLEASE.
Thank you for understanding my position.
Before I comment on your comments, I want you to know where my opinions come from. I started in the industry as a translator (en_pt_BR) almost 30 years ago. I worked for big localization clients that hired the largest big LSPs. I worked for the largest LSPs. I go to the ATA conferences often and try to be up-to-date with the industry trends. I saw the birth of Trados and fuzzy matches, which were a huge disruption for the industry and we adapted. What is a 70%, 90% match worth? Professionals that started when TM was the norm, probably never gave it a second thought. This is why I'm reluctant to take a stand against AI related jobs. I've seen AI training jobs that seem quite interesting for language professionals, and I'd do it if the compensation was what I'd think is fair and gave me an acceptable living wage.
Another point we need to keep in mind is that ProZ.com is not a body able to dictate or influence the market; as you said ProZ.com is not an union and our mission is to help you achieve your goals. Your decisions are what influence and dictate the market. ProZ.com is the aggregate of its members and its perceived "power" depends solely on the group. We give our members the means, like forums and the ability to reach out to other language professionals, so you can get together, discuss and agree on a course of action. Some users don't think that contributing to the existence of ProZ.com is worth the fee, but the power of the community is on their size, not on the ProZ.com brand.
About adapting: translators can't stop the industry transformation but they can shape the profession's future. Adapting doesn't mean accepting low rates, it means understanding what the task is, the effort required to complete it and what the fare rate is. This is what I think translators need to put their energy in. By just ignoring, you might be missing the opportunity to set the rates and "shape the future." Don't forget that clients don't want to train their machine with crap content, so if professionals start responding to the posts saying I cannot do it for $xx but can do it for $yy, it will only take so long for clients to raise their bottom.
Content that human translation is the right/best option will continue to exist, though MT might be adequate for the bulk of the content to translate. With that said, when we get a MTPE job (paying less) and deliver a translation that is no different from a human translation, we are indicating to the clients that it is OK to use MT for content they thought would be better to use human translation. Thus, it is important to rethink what top quality means for each type of translation, not to encourage clients to use MT for everything, since the end result of MTPE is the same as HT.
Clients/LSPs consider edit distance when setting the rate based on the market rate for a no-match per language pair. If individuals think it is too low to do the job, there is a good chance there is a misalignment between what the translator thinks they need to do and the client's expectation (but it is possible that the client wants to push the price down). Keep in mind that as translators compete for the jobs, clients compete for good translators. If they push rates down too hard, they lose. It is true that a buyer, regardless of service/product will always want to pay the least possible, but they also understand that lower cost service/product is not the best or good at all, and at the end they will lose. LSPs will pay more to a translator if they can save time and money in QA steps, for instance.
With all that said, consider opening a forum for your language pair and discuss rates. Encourage your colleagues to enter their rates into their profile and make it not visible, if they prefer. By doing this, you set the market rate for the clients. There will always be the ones who will charge/accept less but a minority is not enough to handle the volume of the industry, so they will not set the rate low.
I'll give you an update about my conversation with our team on a second post.
| | |
|
|
|
| Update on your ideas | Oct 16, 2024 |
I took your ideas to the team, here is an update
1) Make AI related jobs more distinct: This is on the making in the context of the job post form update. There are other services that some language professionals are interested, like subtitling, voice over, that others are not. So we need to address this too.
2) The fair market badge: I took the idea to the team, they scratched their heads and we started a discussion about how to realize that; it seems to be complex, but... See more I took your ideas to the team, here is an update
1) Make AI related jobs more distinct: This is on the making in the context of the job post form update. There are other services that some language professionals are interested, like subtitling, voice over, that others are not. So we need to address this too.
2) The fair market badge: I took the idea to the team, they scratched their heads and we started a discussion about how to realize that; it seems to be complex, but nobody rejected it, so we will continue to discuss it. Keep the ideas coming.
They reminded me about a feature that ProZ.com has that you might not be aware of because you don't search for translators . When clients enter a budget during their search, they are notified if the rate is below community rate. If they keep it, they will get fewer results. See the orange block in the screenshot.
The catch is, if the rates translators entered are outdated or inflated, or the translators don't enter any, this feature doesn't work as well. We would have to carry a campaign to get people to enter/update their rate (can make it invisible to the public) and create a rate category for AI training, for instance, per task category.
Something to think about: if we have an up-to-date "community rate," we could use this feature and not show the AI posts or any job type with low rate offer to professionals who choose to honor the community rate specific of each task type, and not accept lower rates. ▲ Collapse
| | | | Jana T.
Espagne
Local time: 00:09
Membre (2021)
anglais vers espagnol
+ ...
| Another suggestion | Oct 17, 2024 |
Monica Oliveira wrote:
The catch is, if the rates translators entered are outdated or inflated, or the translators don't enter any, this feature doesn't work as well. We would have to carry a campaign to get people to enter/update their rate (can make it invisible to the public) and create a rate category for AI training, for instance, per task category.
Something to think about: if we have an up-to-date "community rate," we could use this feature and not show the AI posts or any job type with low rate offer to professionals who choose to honor the community rate specific of each task type, and not accept lower rates.
I think this would be a good solution.
As some other members have already said, even if I agree that these offers don't belong here, I also believe that banning this kind of jobs won't solve the problem.
No, most of them aren't language-related jobs, but they'll keep posting them and lying in their descriptions. So the system of filters Monica just suggested could help:
AI training as a category to add to/remove from your profile.
A desired rate that is invisible to the public.
Edit: soooo I had a fantastic idea that I suggested here just to go to the job board and see it already exists (I just happen to rarely use it).
It's about the "Delete offer" button. I've just suggested the already existing "Delete offer" button —apparently, my Mediterranean brain doesn't work well when it rains
But what about letting other professionals see the reason why other members have deleted the offer? Just as we have the Blue Board to see how people rate companies, maybe we can see how many people rejected an offer due to low rates or irrelevance. This should be anonymous, of course. But maybe this could encourage new, inexperienced freelancers not to accept jobs that offer humiliating rates.
[Editat el 2024-10-17 10:59 GMT]
[Editat el 2024-10-17 11:17 GMT]
| | | | | Delete offer button | Oct 17, 2024 |
Jana T. wrote:
It's about the "Delete offer" button. I've just suggested the already existing "Delete offer" button —apparently, my Mediterranean brain doesn't work well when it rains
But what about letting other professionals see the reason why other members have deleted the offer? Just as we have the Blue Board to see how people rate companies, maybe we can see how many people rejected an offer due to low rates or irrelevance. This should be anonymous, of course. But maybe this could encourage new, inexperienced freelancers not to accept jobs that offer humiliating rates.
[Editat el 2024-10-17 10:59 GMT] [Editat el 2024-10-17 11:17 GMT]
Thanks Jana, for taking the time to check the "Delete offer."
I'm taking to the team and see if it could work for what you're suggesting.
| | | | | Pages sur ce sujet: < [1 2 3 4 5 6] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » About the constant AI training job posts. | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
| | Pastey | Your smart companion app
Pastey is an innovative desktop application that bridges the gap between human expertise and artificial intelligence. With intuitive keyboard shortcuts, Pastey transforms your source text into AI-powered draft translations.
Find out more » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
