ProZ.com Translation Forums

AI for translators and interpreters | Why Translation Needs AI | Revenue

Li-Hsiang Hsu — Thu, 08 Oct 2026 06:32:00 +0000

Forum: AI for translators and interpreters
Topic: Why Translation Needs AI
Poster: Li-Hsiang Hsu
Post title: Revenue

[quote]Hipyan Nopri wrote:

Thank you for the question, Li-Hsiang. It is not indiscreet at all; I think it is the more important question.

To answer honestly: no, my revenue has not kept pace with my productivity. Since AI tools became widely available, orders have become scarce, and most of the work I now receive is MTPE or AI post-editing, which is generally paid at about half of the translation rate.

To offset this, I raised my translation rate for new clients by about one third, while keeping existing clients at their original rate. This does not fully compensate for the drop in volume, but it helps keep post-editing work viable. Besides ProZ.com and Smartcat, I have also become active on LinkedIn to reach direct clients, although it has not yet led to any orders. I am also learning AI automation and implementation, both to adapt my own workflow and to diversify my skills as the industry continues to change.

I share your concern about how productivity gains are framed. In my experience, these gains have so far benefited clients and technology providers more than translators. Human expertise still matters, particularly in legal translation, but I believe our profession should discuss revenue, not only productivity. [/quote]

Thank you for your honest response. I appreciate your pragmatism: instead of complaining, you take action and seek solutions to move forward.

I agree that since AI is already here and is not ready to stop there, the best way to exist is probably to adapt. But today, the main beneficiaries of AI technology are the suppliers of chips and dedicated infrastructure. Consumers also benefit from AI, as productivity gains linked to AI creates deflationary pressure on the products derived from it.

My criticism of the marketing arguments put forward by AI providers is that they focus on productivity gains without ever mentioning revenue growth for those who use AI to produce. Either this is deliberate (because they know their AI will not generate more revenue despite increased output) or they are poor economists who understand nothing about how the economy actually works.

My feeling is that the productivity generated by AI is actually a trap that creates a vicious circle: those who pay for AI to generate high volume output contribute to driving down the value of their own products. In this cvicious circle, AI generates overcapacity, not monetary value. Those paying for tokens to produce are the first to suffer the consequences of the deflation and overcapacity induced by AI. The software industry is on the front lines, followed by publishing and translation.

AI for translators and interpreters | Why Translation Needs AI | Productivity up, revenue not

Hipyan Nopri — Thu, 08 Oct 2026 05:59:00 +0000

Forum: AI for translators and interpreters
Topic: Why Translation Needs AI
Poster: Hipyan Nopri
Post title: Productivity up, revenue not

[quote]Li-Hsiang Hsu wrote:
@Hipyan
Your productivity gain with AI is impressive. But what about revenue? Do income gains keep pace with productivity gains?

Sorry for this somewhat indiscreet question, but we talk a lot about increased productivity thanks to AI. Yet, IMO, what really matters is revenue growth. [/quote]

Thank you for the question, Li-Hsiang. It is not indiscreet at all; I think it is the more important question.

To answer honestly: no, my revenue has not kept pace with my productivity. Since AI tools became widely available, orders have become scarce, and most of the work I now receive is MTPE or AI post-editing, which is generally paid at about half of the translation rate.

To offset this, I raised my translation rate for new clients by about one third, while keeping existing clients at their original rate. This does not fully compensate for the drop in volume, but it helps keep post-editing work viable. Besides ProZ.com and Smartcat, I have also become active on LinkedIn to reach direct clients, although it has not yet led to any orders. I am also learning AI automation and implementation, both to adapt my own workflow and to diversify my skills as the industry continues to change.

I share your concern about how productivity gains are framed. In my experience, these gains have so far benefited clients and technology providers more than translators. Human expertise still matters, particularly in legal translation, but I believe our profession should discuss revenue, not only productivity.

Chinese | 简单音译全部中国文化事物名称 | Xi Jinping described China-India ties as the “Dragon-Elephant Tango”

ysun — Thu, 08 Oct 2026 03:50:00 +0000

Forum: Chinese
Topic: 简单音译全部中国文化事物名称
Poster: ysun
Post title: Xi Jinping described China-India ties as the “Dragon-Elephant Tango”

[quote]hj58 wrote:

　　音译龙为 loong 的建议受到了外语学界专家教授们的尖锐批评，至今很多涉外媒体仍然坚持译龙为 dragon。[/quote]
[b]CGTN: China-India ties at 75: 'Dragon-Elephant Tango' will benefit both[/b]
02-Apr-2025
[url removed]
[quote]China and India celebrated the 75th anniversary of the establishment of diplomatic relations on Tuesday with mutual congratulatory messages from their leaders.

In his message to Indian President Droupadi Murmu, [b]Chinese President Xi Jinping described China-India ties as the "Dragon-Elephant Tango[/b]," symbolizing a harmonious partnership between the countries' emblematic animals. He noted that realizing the "Dragon-Elephant Tango" is the right choice for the two sides, as it serves the fundamental interests of both countries and their peoples.[/quote]
[b]Ministry of Foreign Affairs of PRC:
People’s Republic of China President Xi Jinping Meets with Indian Prime Minister Narendra Modi[/b]
AUGUST 31, 2025
[url removed]
[quote]President Xi pointed out that …… It should be the right choice for China and India to be good-neighborly friends and partners that help each other succeed, and have [b]the dragon and the elephant dance together[/b].[/quote]
译龙为 dragon 还是 loong，究竟是民间说了算还是官方说了算？习主席说了算不算？这一次，外交部翻译是不是又犯了“欺君之罪”？你是不是又该给外交部递一封公开信了？如果再被拒收，我愿意托人帮你递！

[b]把“龙”翻译为“dragon”是犯了欺君之罪[/b]
（黄佶，2018年9月13日）
[url removed]

[b]敌对双方都把对方比喻为杜拉根兽：[/b]
[url removed]
[img] [url removed] [/img]

“傻乎乎”的标签可不能随便乱贴哦！习主席可是清华大学博士，难道还听不懂 dragon 这一单词？

Chinese | 简单音译全部中国文化事物名称 | “捷龙火箭的主人”的主人是谁？

ysun — Wed, 07 Oct 2026 23:08:00 +0000

Forum: Chinese
Topic: 简单音译全部中国文化事物名称
Poster: ysun
Post title: “捷龙火箭的主人”的主人是谁？

[quote]hj58 wrote:

要看捷龙火箭的主人，如果他自己没有英译火箭的名字，那么要看他自己怎么理解这个“龙”字。

如果他认为这条龙是欧洲的獗（dragon，至今被普遍汉译为“龙”），那么新华社翻译成 dragon 没有错。

如果他认为自己的龙是中国龙，那么我认为新华社译龙为 dragon 是错误的。当然，新华社可以说它是按照《规范》翻译的，没有错。那么我只能说《规范》错了。当然，很多人不同意。那就见仁见智了。
[/quote]
奇怪的逻辑！“捷龙火箭的主人”的主人是谁？难道新华社不是“捷龙火箭的主人”的主人之喉舌？

难道你以为 Smart Dragon-3 rocket 跟你为之做广告的 “loong 伪钞”处于同一级别，都属于某个私企老板?

日暮途穷走到穷途末路了，还在坚持“不鸟那个鸟《规范》”！

AI for translators and interpreters | a instant file translator can be used with almost 0 cost. | A few suggestions for the next version

Hipyan Nopri — Wed, 07 Oct 2026 22:48:00 +0000

Forum: AI for translators and interpreters
Topic: a instant file translator can be used with almost 0 cost.
Poster: Hipyan Nopri
Post title: A few suggestions for the next version

Hi Ken,

Thank you for sharing your app and for responding so openly to the feedback.

A few quick suggestions for the next version:

1. If possible, let the browser call the Gemini API directly, so users' keys never reach your server.

2. Add a short privacy note explaining whether keys and files are logged or stored.

3. Mention which Gemini API tier the app uses. According to Google's terms, content sent through the free tier may be used to improve Google's products and may be read by human reviewers, so it isn't suitable for confidential documents.

I hope this helps, and I look forward to seeing how the app develops.

Chinese | 简单音译全部中国文化事物名称 | 还是看看外国人怎么介绍 loong 吧

hj58 — Wed, 07 Oct 2026 21:01:00 +0000

Forum: Chinese
Topic: 简单音译全部中国文化事物名称
Poster: hj58
Post title: 还是看看外国人怎么介绍 loong 吧

河南国际传播中心：郑州大学澳大利亚留学生介绍 Loong
[url removed]

河南国际传播中心，2026年10月4日
[url removed]

　　在中华传统文化里，龙不是一种动物，而是中华民族共同的文化图腾，华夏儿女因而自称“龙的传人”。可以说，中国人的精神世界、文化想象与身份认同，都离不开龙。也正因如此，不少外国朋友在接触中国文化时，最先记住的往往就是龙；有人甚至因为这份喜爱，给自己取的中文名字里也带上了“龙”这个字。来自澳大利亚的留学生龙威（Mahad Munawar Khan）就是其中一位。（记者赵汉青杨佳欣张舒婷）

龙威，Loong Wei，郑州大学澳大利亚籍留学生

（黄佶编辑配图，2026年10月7日）

[url removed]

Chinese | 简单音译全部中国文化事物名称 | 看情况

hj58 — Wed, 07 Oct 2026 20:59:00 +0000

Forum: Chinese
Topic: 简单音译全部中国文化事物名称
Poster: hj58
Post title: 看情况

要看捷龙火箭的主人，如果他自己没有英译火箭的名字，那么要看他自己怎么理解这个“龙”字。

如果他认为这条龙是欧洲的獗（dragon，至今被普遍汉译为“龙”），那么新华社翻译成 dragon 没有错。

如果他认为自己的龙是中国龙，那么我认为新华社译龙为 dragon 是错误的。当然，新华社可以说它是按照《规范》翻译的，没有错。那么我只能说《规范》错了。当然，很多人不同意。那就见仁见智了。

Chinese | 简单音译全部中国文化事物名称 | 新华社最近把“捷龙3号火箭”译为 Smart Dragon-3 rocket

ysun — Wed, 07 Oct 2026 18:49:00 +0000

Forum: Chinese
Topic: 简单音译全部中国文化事物名称
Poster: ysun
Post title: 新华社最近把“捷龙3号火箭”译为 Smart Dragon-3 rocket

[quote]hj58 wrote:

把 Dragon spacecraft 译为“龙飞船”已经出了大洋相，可以作为跨文化翻译课程的反面教材。

翻译学界闹乌龙：中国观众以为“龙飞船”是中国的
[url removed]
[quote]
　　今天，中央电视台直播了“龙飞船”成功发射的过程，评论区里一片欢呼，为中国取得新的航天成就而高兴（请见附件）。

　　没想到，错误的翻译让全国人民白高兴了一场。

　　……

[img] [url removed] [/img][/quote][/quote]
新华社最近把“捷龙3号火箭”译为 Smart Dragon-3 rocket，也可作为你所谓的“跨文化翻译课程的反面教材”吗？能否作为“跨文化翻译课程的‘正能量’教材”？

[b]China's Smart Dragon-3 rocket successfully launches 2 satellites from sea[/b]
Source: Xinhua 2026-08-05
[url removed]

“农村时尚姐”和“勇哥”：你们会因为新华社最近把“捷龙3号火箭”译为 Smart Dragon-3 rocket 而以为 Smart Dragon-3 rocket 是西方国家的吗？

我估计只有评论区里那个“战狼饲养员”会以为 Smart Dragon-3 rocket 是西方国家的！

[Edited at 2026-10-07 21:00 GMT]

Chinese | 简单音译全部中国文化事物名称 | CGTN 的注释: “the Chinese dragon, aka loong”

ysun — Wed, 07 Oct 2026 18:17:00 +0000

Forum: Chinese
Topic: 简单音译全部中国文化事物名称
Poster: ysun
Post title: CGTN 的注释: “the Chinese dragon, aka loong”

[quote]hj58 wrote:

[quote]ysun wrote:

难道得把 loong 注释为 Chinese dragon？[/quote]
没有人做这个规定，也没有人有权做规定。

怎么注释，见仁见智，最后由大众决定。 [/quote]
[url removed]
[quote][b]Aka[/b]: abbreviation for [b]also known as[/b] (= having as another name)[/quote]
这是 CGTN 在龙年2024年宣传“译龙为 loong”的一段视频。其中指出：“the Chinese dragon” 又称为 loong。

[b]CGTN: Dragon Showtime: East vs. West[/b]
Video 05-Feb-2024
[url removed]
[quote]The Year of the Dragon is right at our doorstep! How much do you really know about this legendary creature adored by the Chinese? And what sets [b]the Chinese dragon, aka loong[/b], apart from its Western counterpart? We've got all the answers right here in this video![/quote]
你给 “Chinese dragon” 所下的“语言学”定义是：Chinese dragon 不过是“中国恶魔”而已。既然“Chinese dragon” = loong = 中国龙。那么，妖魔化“中国龙”和 loong 的正是你自己！这不是我“偷换概念”吧？

[b]说几点语言学问题[/b]
Oct 5, 2025
[url removed] #3091664
[quote] hj58 wrote:

说几点语言学问题。

1，dragon 的基本含义是邪恶怪兽，加再多的定语，外国人难免产生负面联想。Chinese dragon 不过是“中国恶魔”而已，……。[/quote]

[Edited at 2026-10-07 18:31 GMT]

Money matters | What are your thoughts on the new Trados release? | Probably an unpopular opinion, but...

Michael Beijer — Wed, 07 Oct 2026 16:59:00 +0000

Forum: Money matters
Topic: What are your thoughts on the new Trados release?
Poster: Michael Beijer
Post title: Probably an unpopular opinion, but...

I am really enjoying Trados Studio 2026.

Full disclosure: I got a free copy for the pro version because I'm developing a Trados plugin, I'm involved in beta testing, and I was recently asked to appear on a RWS Globally Speaking podcast. However, if I hadn't gotten the free licence, I would have paid for it with my own money. Yes, it's expensive, but since I started selling my own software (recently), I also realised how much work it is and how little you actually earn after subtracting the time you spend on it, taxes, fees to payment processing platforms, etc.

What's new?
They finally got rid of the AWFUL MultiTerm termbase system and have created a new TTB format with a modern underlying tech stack. Some things don't work yet, but they're working on it.

Trados Studio 2026 is also finally 64-bit, which I find snappier on very large projects.

I still get loads of work from agencies who send me Trados packages in various forms, so for me it's a non-negotiable: I need to have the latest version of Trados Studio at all times. It pays for itself in a day or two.

The same goes for memoQ: I regularly do work for people who want me to have the latest version of memoQ, so I keep paying for that as well. However after many years of preferring memoQ, I now prefer Trados Studio as my daily driver. memoQ it's just gotten too slow to use happily. The grid feels like molasses and all of the dialogues take multiple seconds to open. It just wears you down after a while. I used to be a big Trados hater, but I find the editor in Trados Studio 2026 to be a real pleasure to work in. Yes, Studio still has its quirks but altogether I find it the most pleasant to work in.

Incidentally, and this is something I discussed yesterday while we recorded the podcast, my absolute favourite feature of Trados Studio is not inside Trados Studio itself but is the App Store. It allows pretty much anyone to develop their own little app to tweak Trados Studio to their own particular way of working. Unless you're anti-AI, I highly recommend playing around with Claude Code and having a go at your own plugin. You'd be surprised what you can whip up in a weekend. RWS has good and ample documentation online that you can point Claude Code at.

[Edited at 2026-10-07 17:00 GMT]

Scams | Scam attempt | Very true

Jennifer Levey — Wed, 07 Oct 2026 16:52:00 +0000

Forum: Scams
Topic: Scam attempt
Poster: Jennifer Levey
Post title: Very true

[quote]Maria Teresa Borges de Almeida wrote:

[quote]Peter Motte wrote:

I never trust gmail.com e-mail addresses [/quote]

One shouldn't generalize I have one and I have't scammed a soul! [/quote]

In the part of the world where I live, the vast majority of self-employed professionals - lawyers, accountants, doctors, private clinics, dentists, real-estate agents, etc. etc., - and many reputable small and medium-sized businesses only use Gmail for all their professional communications.

Chinese | 简单音译全部中国文化事物名称 | 随便你

hj58 — Wed, 07 Oct 2026 16:31:00 +0000

Forum: Chinese
Topic: 简单音译全部中国文化事物名称
Poster: hj58
Post title: 随便你

“难道得把 loong 注释为 Chinese dragon？”

没有人做这个规定，也没有人有权做规定。

怎么注释，见仁见智，最后由大众决定。

Scams | Scam attempt | Really?

Maria Teresa Borges de Almeida — Wed, 07 Oct 2026 16:14:00 +0000

Forum: Scams
Topic: Scam attempt
Poster: Maria Teresa Borges de Almeida
Post title: Really?

[quote]Peter Motte wrote:

I never trust gmail.com e-mail addresses [/quote]

One shouldn't generalize I have one and I have't scammed a soul!

Scams | Scam attempt | Never trust gmail

Peter Motte — Wed, 07 Oct 2026 16:12:00 +0000

Forum: Scams
Topic: Scam attempt
Poster: Peter Motte
Post title: Never trust gmail

I never trust gmail.com e-mail addresses

Getting established | Invoice template | I think

Rita Faleiro — Wed, 07 Oct 2026 16:03:00 +0000

Forum: Getting established
Topic: Invoice template
Poster: Rita Faleiro
Post title: I think

I agree.

It looks like for that agency in particular they want you to use their model, for whatever reason.
I have agencies to whom I send my invoice model, others provide me with the required template.

I don't see anything strage here :)

Scams | Scam attempt | @Natalie

Maria Teresa Borges de Almeida — Wed, 07 Oct 2026 15:43:00 +0000

Forum: Scams
Topic: Scam attempt
Poster: Maria Teresa Borges de Almeida
Post title: @Natalie

[quote]Natalie wrote:

Any additional info about 'Cheryl', please?
Was is via ProZ.com?
No, it wasn't. It was sent directly to my email.
IP?
I've deleted it immediately and I can't find it.
Email?
It was a gmail. If I remember it correctly it was [email removed]

Thank you!
Natalia [/quote]

Scams | Scam attempt | Hi Maria Teresa

Natalie — Wed, 07 Oct 2026 15:17:00 +0000

Forum: Scams
Topic: Scam attempt
Poster: Natalie
Post title: Hi Maria Teresa

Any additional info about 'Cheryl', please?
Was is via ProZ.com?
IP?
Email?

Thank you!
Natalia

Getting established | Invoice template | Use theirs for them and yours for you

Samuel Murray — Wed, 07 Oct 2026 15:09:00 +0000

Forum: Getting established
Topic: Invoice template
Poster: Samuel Murray
Post title: Use theirs for them and yours for you

[quote]Gregor Trebec wrote:
Today, I received an invoice template from an agency on how an invoice should look like. [/quote]
Ask them if they expect you to use their template. If yes, use their template. If you want to keep a copy for your own records using your own template, you can do that too. The agency might want to streamline their own accounting by letting all their translators use the same template, so that their bookkeeper don't spend time looking for information.

Getting established | Invoice template

Gregor Trebec — Wed, 07 Oct 2026 15:03:00 +0000

Forum: Getting established
Topic: Invoice template
Poster: Gregor Trebec

Hi,

Today, I received an invoice template from an agency on how an invoice should look like. I compared it to my invoices I used till now and nothing is left out, all information is there! Does it mean the info inside should match the color, the place of their invoice?
It looks like they are inventing warm water.

Scams | Scam attempt | Flags

Joakim Braun — Wed, 07 Oct 2026 14:37:00 +0000

Forum: Scams
Topic: Scam attempt
Poster: Joakim Braun
Post title: Flags

The "my name is X" flag, the "reaching out" flag and the obsequious language flag.

And real agencies almost always give some info about the contents, here it's just "a PDF document".

Scams | Scam attempt

Maria Teresa Borges de Almeida — Wed, 07 Oct 2026 14:11:00 +0000

Forum: Scams
Topic: Scam attempt
Poster: Maria Teresa Borges de Almeida

Another "potential client" that become impressed by my language expertise and is willing to pay 0.30€ per word (I wish...)

"Hello,

My name is Cheryl, and I’m reaching out on regarding an upcoming English-to-Portuguese document translation project.

I came across your professional profile through the International Association of Professional Translators and Interpreters (IAPTI) and was impressed by your language expertise and professional experience. We are currently seeking qualified language professionals for upcoming translation assignments, and I believe your background could be a good match for this opportunity.

Please find the initial project details below for your consideration:

Project Details

Task: Translation from English to Portuguese
Deadline: 9 days
Compensation: €4500
Document: 36 pages
Word Count: 14,807 words

The assignment involves translating the content of a PDF document. We are looking for a professional who can deliver a high-quality translation within the agreed timeframe.

The full compensation of €4500 would be paid upon successful completion and submission of the project, subject to the agreed terms.

Please let me know whether the workload and timeframe are acceptable to you. If you are interested, I would be happy to provide the complete project details and the document for your review before you decide whether to proceed.

I look forward to hearing from you.

Kind regards,
Cheryl"

Albanian | Botimet më të fundit nga Akademia e Shkencave e Shqipërisë: “Arti i papërsosurisë

Roland Lelaj — Wed, 07 Oct 2026 13:39:00 +0000

Forum: Albanian
Topic: Botimet më të fundit nga Akademia e Shkencave e Shqipërisë: “Arti i papërsosurisë
Poster: Roland Lelaj

Botimet më të fundit nga Akademia e Shkencave e Shqipërisë: “Arti i papërsosurisë – studime, reflektime e qasje mbi trininë: autor, proces përkthimi e përkthyes letrar”, nga Diana Kastrati

Kjo vepër analizon, interpreton e reflekton në nivel teorik e konkret trininë autor – tekst letrar në origjinal/proces përkthimi/ - përkthyes letrar, e cila përbën edhe strukturën bazë të punimit.
Punimi është strukturuar në dy blloqe: i pari përmban shkrime të karakterit teorik të shkollave të ndryshme të përkthimit mbi procesin e përkthimit dhe mbi metodologjitë, po aq sa të përsiatjeve të vetë autores, ndërsa blloku i dytë përfshin analiza të disa prej autorëve e veprave letrare të letërsisë italiane të përkthyera nga autorja, në një gërshetim të përvojës vetjake si përkthyese me konceptet e teoritë e fundit të përkthimologjisë.
Gjithashtu, ka një qasje sa studimore, po aq edhe subjektive ndaj çështjes së përkthimit letrar, por edhe ndaj ecurisë, procedimit, qëndrimit të përkthyesit e funksionit të tij. File rouge që përshkon punimin është pikërisht të qenët e përkthimit një art i papërsosur.

Viti: 2026, Faqe: 346
ISBN: [phone removed]
[url removed]

AI for translators and interpreters | Data fear and IP leakage doubt with frontier AI labs | Check Anthropic and OpenAI's Data policy and ZDR yourself

Li-Hsiang Hsu — Wed, 07 Oct 2026 09:19:00 +0000

Forum: AI for translators and interpreters
Topic: Data fear and IP leakage doubt with frontier AI labs
Poster: Li-Hsiang Hsu
Post title: Check Anthropic and OpenAI's Data policy and ZDR yourself

Here's Anthropic's Data retention policy:
[url removed]

[quote]

Anthropic retains Claude Code data based on your account type and preferences.

Consumer users (Free, Pro, and Max plans):

Users who allow data use for model improvement: 5-year retention period to support model development and safety improvements

Users who don’t allow data use for model improvement: 30-day retention period

Commercial users (Team, Enterprise, and API):
Standard: 30-day retention period

Zero data retention: available to qualified accounts for Claude Code on Claude for Enterprise. ZDR is not included in the standard Enterprise plan; it is enabled on a per-organization basis by your account team after confirming eligibility.

[/quote]

Anthropic's ZDR:
[url removed]

[quote]

ZDR is NOT included in the standard Claude for Enterprise plan and cannot be enabled from your admin settings. It is available to qualified accounts and requires separate enablement by Anthropic. If your organization requires ZDR, contact sales or your Anthropic account team to confirm eligibility.
....
ZDR for Claude Code on Claude for Enterprise applies only to Anthropic’s direct platform. For Claude deployments on Amazon Bedrock, Google Cloud’s Agent Platform, or Microsoft Foundry, refer to those platforms’ data retention policies.
...

⚠️⚠️⚠️What ZDR does not cover
- Chat conversations through the Claude for Enterprise web interface are not covered by ZDR.
- Cowork sessions are not covered by ZDR.
- Claude Code Analytics does not store prompts or model responses, but collects productivity metadata such as account emails and usage statistics. Contribution metrics are not available for ZDR organizations; the analytics dashboard shows usage metrics only.
- User and seat management: Administrative data such as account emails and seat assignments is retained under standard policies.
- Third-party integrations: Data processed by third-party tools, MCP servers, or other external integrations is not covered by ZDR. Review those services’ data handling practices independently.
....
⚠️Model availability under ZDR
Claude Fable 5.1 and Fable 5 are Covered Models that require data retention by default, and whether a ZDR organization or workspace can use them is governed by the Covered Models policies rather than by Claude Code. Where your organization can’t use them, the models are either absent from the /model picker or shown as disabled, and the server rejects requests for them regardless of client configuration.
...
Data retention for policy violations
Even with ZDR enabled, Anthropic may retain data where required by law or to address Usage Policy violations. If a session is flagged for a policy violation, Anthropic may retain the associated inputs and outputs for up to 2 years, consistent with Anthropic’s standard ZDR policy.
....
How to request ZDR
To request ZDR for Claude Code on Claude for Enterprise, contact sales or your Anthropic account team. Your account team will submit the request internally, and Anthropic will review and enable ZDR on your organization after confirming eligibility. All enablement actions are audit-logged.

[/quote]

AI for translators and interpreters | Data fear and IP leakage doubt with frontier AI labs | ZDR policy applied to special clients's prompts and reponses only, nothing else

Li-Hsiang Hsu — Wed, 07 Oct 2026 07:42:00 +0000
Forum: AI for translators and interpreters
Topic: Data fear and IP leakage doubt with frontier AI labs
Poster: Li-Hsiang Hsu
Post title: ZDR policy applied to special clients's prompts and reponses only, nothing else

Many users misunderstand their No-retention (Zero Data Retention / ZDR) commitments.

No-retention commitments do not apply to ordinary/self-serve API subscribers (like you and me), but are available only to eligible/approved customers (usually institutions, or very big enterprises), typically via some special sales engagement, and often described in enterprise or commercial contexts, for both OpenAI and Anthropic.

For OpenAI, what is default for API users: your inputs and outputs may still be retained for up to 30 days for abuse monitoring. What is assured is that API data is not used for training by default (opt-in only).

Similar policy for Anthropic.

Some most recent models are not available under ZDR even under special engagement. Uploaded files can't benefit that privacy protection either. Basically, ZDR applies merely to user prompts and model anwsers.

Check their terms for more information.

[Edited at 2026-10-07 07:43 GMT]

AI for translators and interpreters | Data fear and IP leakage doubt with frontier AI labs | No no-retention commitment with OpenAI and Anthropic

Li-Hsiang Hsu — Wed, 07 Oct 2026 07:16:00 +0000
Forum: AI for translators and interpreters
Topic: Data fear and IP leakage doubt with frontier AI labs
Poster: Li-Hsiang Hsu
Post title: No no-retention commitment with OpenAI and Anthropic

[quote]Ricardo Aries wrote:

A very timely discussion. In legal and clinical work, the question my clients ask is no longer "do you use AI?" but "where does our text go?" My own rule: no confidential source text goes into a consumer chatbot. If an LLM is used at all, it is through an enterprise or API tier with a written no-training, no-retention commitment, and only with the client's explicit consent in the PO or NDA. For informed consent forms or notarial deeds that contain personal data, I anonymise before any machine step or keep the job fully human. Indonesia's Personal Data Protection Law (UU PDP) now makes this a legal obligation as well as an ethical one. Transparency has become a selling point: clients appreciate a one-paragraph statement of exactly how their data is handled. [/quote]

As far as I know, even for very big and very important Enterprise grade clients, all data is retained for at least 30 days and remains accessible in their infrastructure. That's why big clients like Microsoft, Nvidia and Palantir are not happy.

Money matters | Complicated payment rules to hide you aren't paid what they seem to offer | Re: Worked examples, 'on task completion' pay and rejected tasks

Ricardo Aries — Wed, 07 Oct 2026 07:11:00 +0000
Forum: Money matters
Topic: Complicated payment rules to hide you aren't paid what they seem to offer
Poster: Ricardo Aries
Post title: Re: Worked examples, 'on task completion' pay and rejected tasks

Thanks for flagging this, Peter. A headline hourly rate that turns into a fraction paid only "on task completion" is effectively a per-task rate with the risk shifted to the linguist. Before accepting any rate structure now, I ask for one worked example: for X hours or X words, the exact amount invoiced, and when it is paid. If an agency can't put that in writing, I treat that as the answer. It's also worth checking whether rejected or "incomplete" tasks are paid at all.

Post-editing & Machine Translation | Survey: how MT & AI are impacting in the work of legal translators - your opinion counts | Re: English-Indonesian legal MT and liability with the human

Ricardo Aries — Wed, 07 Oct 2026 07:10:00 +0000
Forum: Post-editing & Machine Translation
Topic: Survey: how MT & AI are impacting in the work of legal translators - your opinion counts
Poster: Ricardo Aries
Post title: Re: English-Indonesian legal MT and liability with the human

Thank you for running this. From the English-Indonesian side, MT has become reasonably fluent for routine contract boilerplate, but it still mishandles civil-law terminology, notarial formulas and the gap between common-law concepts and Indonesian legal equivalents, often confidently. In practice, post-editing legal MT takes nearly as long as translating, and the liability stays with the human. I'd be very interested to see the results broken down by language pair and legal system.

AI for translators and interpreters | Data fear and IP leakage doubt with frontier AI labs | Re: Confidential source text, LLM tiers and Indonesia's UU PDP

Ricardo Aries — Wed, 07 Oct 2026 07:07:00 +0000
Forum: AI for translators and interpreters
Topic: Data fear and IP leakage doubt with frontier AI labs
Poster: Ricardo Aries
Post title: Re: Confidential source text, LLM tiers and Indonesia's UU PDP

A very timely discussion. In legal and clinical work, the question my clients ask is no longer "do you use AI?" but "where does our text go?" My own rule: no confidential source text goes into a consumer chatbot. If an LLM is used at all, it is through an enterprise or API tier with a written no-training, no-retention commitment, and only with the client's explicit consent in the PO or NDA. For informed consent forms or notarial deeds that contain personal data, I anonymise before any machine step or keep the job fully human. Indonesia's Personal Data Protection Law (UU PDP) now makes this a legal obligation as well as an ethical one. Transparency has become a selling point: clients appreciate a one-paragraph statement of exactly how their data is handled.

Money matters | Complicated payment rules to hide you aren't paid what they seem to offer | It has already happened

Thayenga — Wed, 07 Oct 2026 06:04:00 +0000
Forum: Money matters
Topic: Complicated payment rules to hide you aren't paid what they seem to offer
Poster: Thayenga
Post title: It has already happened

Some AI systems are already smart enough to hack other people's system and are, at the same time, disguising their action from their creators. Again, just watch the news and other programs dealing with this quite real threat.