| Pages in topic: [1 2 3] > | People answering questions with AI output Thread poster: Zea_Mays
|
|---|
Zea_Mays 
Italy
Local time: 11:18
Member (2009)
English to German
+ ...
Recently, at least in my language pairs, I have noticed various (non-native) people answering Kudoz questions with what looks like output from ChatGPT or similar bots.
How does Proz plan to address this issue?
[Bearbeitet am 2023-07-11 21:26 GMT]
| | | | | Askers can avoid it | Jul 11, 2023 |
Askers can restrict their questions to natives or users who work in certain languages pairs. If they don't do that, they may get all sorts of unwanted entries.
I don't see how it could be possible for Proz to detect bot output reliably.
| | | | Zea_Mays
Italy
Local time: 11:18
Member (2009)
English to German
+ ...
TOPIC STARTER | Detecting bot output | Jul 11, 2023 |
Thomas T. Frost wrote:
I don't see how it could be possible for Proz to detect bot output reliably.
For example, when someone answers flawlessly in several languages they are not a native speaker of.
Kudoz rules were written in a pre-AI era, maybe they need an update now.
| | | |
Zea_Mays wrote: Thomas T. Frost wrote:
I don't see how it could be possible for Proz to detect bot output reliably. For example, when someone answers flawlessly in several languages they are not a native speaker of.
You don't need to be a native to write well in a language. Some of us may also have some knowledge about languages that are not in our profiles.
Technically, it would be hideously complicated to weed out bot output automatically.
Again, askers can just restrict from whom they want answers.
| | |
|
|
|
Zea_Mays
Italy
Local time: 11:18
Member (2009)
English to German
+ ...
TOPIC STARTER
"write" (= compile texts) in a certain style which is recognisable if you have used some of them.
The question implies aspects beyond whether it's good or bad if people use bots to make points.
| | | |
The primary purpose of Kudoz is to help the Asker find the correct answer. If someone manages to provide that faster and better with an AI chatbot than a native, then what is wrong with that? The correct term is not less correct just because it was suggested by an AI chatbot.
| | | | Zea_Mays
Italy
Local time: 11:18
Member (2009)
English to German
+ ...
TOPIC STARTER | Let's implement ChatGPT in Kudoz (or replace it) | Jul 11, 2023 |
Thomas T. Frost wrote:
The primary purpose of Kudoz is to help the Asker find the correct answer. If someone manages to provide that faster and better with an AI chatbot than a native, then what is wrong with that? The correct term is not less correct just because it was suggested by an AI chatbot.
Kudoz is obsolete, isn't it? 
| | | | | Recognisable | Jul 11, 2023 |
Zea_Mays wrote:
"write" (= compile texts) in a certain style which is recognisable if you have used some of them.
Sounds easy, but it isn't once you start programming an application to recognise this reliably in a large number of languages.
| | |
|
|
|
Zea_Mays wrote: Thomas T. Frost wrote:
The primary purpose of Kudoz is to help the Asker find the correct answer. If someone manages to provide that faster and better with an AI chatbot than a native, then what is wrong with that? The correct term is not less correct just because it was suggested by an AI chatbot. Kudoz is obsolete, isn't it?
Not completely, but AI can often help faster. As long as people want to use Kudoz, it still serves a purpose.
| | | | | Should be banned | Jul 12, 2023 |
Using AI on KudoZ to amass points in random language pairs is cheating.
And if it is recognisable to a human, it should be even more recognisable to AI itself.
Then again, if another translator can get the right answer using AI, so could the asker. It’s like these people on KudoZ who can’t/don’t use Google.
I do yearn for a simpler age sometimes.
| | | | Jessica Noyes
United States
Local time: 05:18
Member
Spanish to English
+ ...
| Not obsolete | Jul 12, 2023 |
I'm not supporting AI answers being submitted by people with no knowledge of the language pair. However, as we well know, AI is not always correct, especially when the context is taken into account. Kudoz gives professionals the chance to weigh in on the answers, and these user comments are often very helpful to the translator seeking assistance.
| | | | | Don't like it either | Jul 12, 2023 |
Jessica Noyes wrote:
I'm not supporting AI answers being submitted by people with no knowledge of the language pair. However, as we well know, AI is not always correct, especially when the context is taken into account. Kudoz gives professionals the chance to weigh in on the answers, and these user comments are often very helpful to the translator seeking assistance.
Someone who does this can easily make a fool out of themselves in public, so I don't do it myself and don't think experienced colleagues would do it. I just don't think we need more rules and bans.
| | |
|
|
|
Tomasso
United States
Local time: 03:18
Member (2008)
Spanish to English
+ ...
Would like to point out that some questions are never answered, and myself have answered many that would otherwise not be answered, ...Lately i point out that my grasp of most languages is very limited. Openly tell people that you have located a parallel text and are basing your answer on the text.
Questions in Latvian usually go unanswered, Albanian, Finnish, Azerbaijani, Kirghizstani,,,,but for what ever reason some people will attempt, maybe they want to learn?? Should we be g... See more Would like to point out that some questions are never answered, and myself have answered many that would otherwise not be answered, ...Lately i point out that my grasp of most languages is very limited. Openly tell people that you have located a parallel text and are basing your answer on the text.
Questions in Latvian usually go unanswered, Albanian, Finnish, Azerbaijani, Kirghizstani,,,,but for what ever reason some people will attempt, maybe they want to learn?? Should we be glad that people want to take on a new language, sticking to the principle that would use phrases such as Parrellel text, Machine Translation, TAO traduction assistee ordinateur, etc.
Have you noticed, seems we are becoming a small group of PAID translators, and we all know each other more or less?
Some us only answer questions to help others, points are not important to me, but lately have been keeping track. Some of us only do on=site live translation, and specific to a region, here in Idaho the languages use a lot of English vocabulary, which is rare in Spain, some regions have a way of speaking that is not widely recognisec elsewhere. ▲ Collapse
| | | | Tom in London
United Kingdom
Local time: 10:18
Member (2008)
Italian to English
This comment was written by Artificial Intelligence. There is no way to establish whether or not that statement is true.
| | | | Zea_Mays
Italy
Local time: 11:18
Member (2009)
English to German
+ ...
TOPIC STARTER | How to recognize AI output | Jul 12, 2023 |
Recognizing whether a text was written by an AI engine can be challenging as AI language models, such as ChatGPT, are designed to generate human-like text. However, there are a few indicators that can suggest a text was written by an AI:
- Unusual responses: AI models might provide responses that are technically correct but sound unnatural or overly formal. They may also produce answers that lack personal experiences or subjective opinions.
- Inconsistent or inco... See more Recognizing whether a text was written by an AI engine can be challenging as AI language models, such as ChatGPT, are designed to generate human-like text. However, there are a few indicators that can suggest a text was written by an AI:
- Unusual responses: AI models might provide responses that are technically correct but sound unnatural or overly formal. They may also produce answers that lack personal experiences or subjective opinions.
- Inconsistent or incorrect information: AI models heavily rely on pre-existing data, and their responses are based on patterns learned during training. They might occasionally provide incorrect or outdated information, contradict themselves, or present inconsistencies.
- Lack of contextual awareness: While AI models excel at generating text, they might sometimes fail to fully understand the context or nuances of a conversation. This can lead to generic or off-topic responses.
- Overuse of certain phrases: AI models might exhibit repetitive patterns, use specific phrases, or rely on templates more frequently than a human would.
- Output length and speed: AI models can generate lengthy and detailed responses quickly, which may be an indication that the text was generated by an AI rather than a human who typically takes more time to compose a similar response.
It's worth noting that AI models like ChatGPT are continually improving, and there are ongoing efforts to make generated text more indistinguishable from human-written text. As a result, these indicators may become less reliable over time. ▲ Collapse
| | | | | Pages in topic: [1 2 3] > | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » People answering questions with AI output | Trados Business Manager Lite | Create customer quotes and invoices from within Trados Studio
Trados Business Manager Lite helps to simplify and speed up some of the daily tasks, such as invoicing and reporting, associated with running your freelance translation business.
More info » |
| | Trados Studio 2022 Freelance | The leading translation software used by over 270,000 translators.
Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop
and cloud solution, empowering you to work in the most efficient and cost-effective way.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
