At a time when almost anyone can upload a document to an artificial intelligence system and receive a translation within seconds, it is easy to see why many people wonder whether professional translators are still needed. The technology is accessible and fast, and in many cases it can produce text that reads naturally and convincingly. On the face of it, the gap between human and automatic translation seems to be narrowing.
Behind the technological question, however, lies a far more fundamental one: what actually makes a translation good? Is it enough for the translated text to be readable and linguistically correct, or must it also preserve the meaning of the original precisely, suit the context in which it will be used, and take into account the possible consequences of choosing one word over another?
Two articles published in the October 2026 issue of ITI Insight, the professional magazine of the Institute of Translation and Interpreting (ITI) in the United Kingdom, offer interesting perspectives on the changes taking place in the field.
The first article
AI, Translation and the Value of Judgement
The first, AI, Translation and the Value of Judgement, by professional translator Lesley Sogl, looks at how the translation process is developing in the age of artificial intelligence and at the importance of human judgement.
The second article
Do Our Clients Trust AI or Us?
The second, Do Our Clients Trust AI or Us?, by Dr Christophe Fricker, Associate Professor of German and Translation at the University of Bristol, examines the meaning of trust in the relationships between clients, translators and technology.
The two articles deal with different aspects of the profession, but together they invite a broader discussion about the value of professional translation in a world where producing text in another language has become an almost instant task.
Translation is not just swapping words
Lesley Sogl’s article, published on pages 14-15 of the magazine, describes how translation work is changing with the arrival of artificial intelligence. The technology can speed up the preparation of first drafts, but at the same time the judgement, expertise and editorial decisions at the heart of good translation are becoming more important.
This is a significant distinction.
Automatic translation systems are no longer limited to simple, literal translation. Advanced systems can recognise linguistic context, suggest alternative wording and produce whole sentences that look as if they were written in the target language in the first place. Sometimes the result is good, and sometimes it can even serve as a useful basis for further professional work.
Even so, the ability to produce a correct sentence does not guarantee that the original meaning has been preserved.
Similarly, translating an academic certificate, a government document or a legal document is not simply a matter of knowing the vocabulary of two languages. It requires an understanding of what the terms mean in the system in which the document was created, and how they can be conveyed accurately into another system.
In Hebrew to English translation, the challenge is particularly significant when Israeli documents are to be submitted in the United Kingdom, the United States or other countries. Institutions with similar names do not necessarily play the same role, and professional or legal qualifications do not always have a direct equivalent.
A professional translator needs to identify these differences, consider the available options and make linguistic decisions according to the context.
In other words, being able to translate is not necessarily the same as being able to decide how something should be translated.
Do we really trust artificial intelligence?
Dr Christophe Fricker’s article, published on pages 24-25, raises an interesting question: does the fact that people use artificial intelligence necessarily mean that they trust it?
Fricker suggests distinguishing between reliance on technology and real trust.
In his words:
“Reliance on technology is not trust.”
To illustrate the difference, Fricker refers to a global study published in 2025, led by researchers at the University of Melbourne in collaboration with KPMG. The study, which included more than 48,000 participants in 47 countries, found that 66% of respondents use artificial intelligence regularly to some extent, but only 46% said they were willing to trust these systems.
48,000+participants
47countries
66%use artificial intelligence regularly to some extent
46%are willing to trust it
The figures do not relate specifically to translation, and on their own they say nothing about the quality of translations produced by artificial intelligence. They do, however, point to an interesting gap between the use of technology and the degree of trust people place in it.
A person may use an artificial intelligence system every day because it saves time, even if they are not entirely sure the result is reliable.
This is especially true when they do not have the knowledge needed to check the information they have received.
In translation, this problem is sometimes even more pronounced.
When someone asks for a document to be translated from a language they do not understand, they may find it hard to spot errors in the final product. A sentence can sound perfectly natural in English, even though its meaning differs from that of the original Hebrew sentence.
In some cases, the very fluency of the translated text can create a sense of confidence that is not justified.
The question, therefore, is not only whether an artificial intelligence system can translate the document, but also how to make sure that the result is faithful to the original and suitable for the purpose for which the translation is needed.
When the client cannot check the translation themselves
One of the distinctive features of translation services is the possible gap in knowledge between the service provider and the client.
When someone orders a service they are able to assess themselves, it is relatively easy for them to judge the result. When they order a translation into a language they do not speak, however, their ability to check the quality of the work is limited.
A client who needs a birth certificate, marriage certificate, academic transcript or legal document translated from Hebrew into English does not always know what the correct translation should look like.
They may not be familiar with the professional terminology, the accepted transliteration conventions or the differences between the administrative and legal systems of different countries.
In these circumstances, the quality of the service is measured not only by the text the client receives, but also by the professional process behind it.
- Do the names match the spelling in the passports?
- Have the dates been copied correctly?
- Have the professional terms been checked in context?
- Have all the essential details of the source document been preserved?
These questions are not about how the translation looks, but about how reliable it is.
A mistranslated name, number, term or date can lead to misunderstandings, delays and even the need to submit documents again.
This is not to say that every automatic translation contains errors or that every human translation is free of them. Human translators can certainly make mistakes, which is why professional review and quality control are an important part of the work.
The difference lies in being able to follow an orderly process in which a person with linguistic expertise reviews the decisions made and takes professional responsibility for the work.
The client is not only looking for a result. They are also looking for confidence
One of the important ideas in Fricker’s article concerns the nature of the relationship between translator and client.
He describes a conversation with a client who wanted to work with human translators because she valued the quality of their work. Yet she felt that dealing with a translation agency was not fundamentally different from using an artificial intelligence tool: you send text and get text back, except that it takes longer.
This example is thought-provoking.
If clients do not feel that there is a real professional relationship between them and the service provider, how can they understand the added value they are receiving?
Fricker stresses that producing a good translation and delivering it to the client is not enough. In his view, the professional relationship needs investment too, not only the final product.
This has very practical implications.
When a client approaches a professional translator, it is sometimes important to establish first what the document is for. Is it for a government authority? A university? An employer? A court? Does the receiving organisation require a certified translation, a translator’s statement, a signature or a document in a particular format?
The client may also not need the whole document translated, only certain pages. In other cases, clarification may be needed about the spelling of names, abbreviations, stamps or handwritten information.
A professional translator can raise these questions before work begins, explain the options and help the client define the service they really need.
This communication is an integral part of the quality of the service.
It also shows why professionalism is not measured only by how quickly a document is translated, but by the ability to understand the client’s needs and act on them.
And what about confidentiality?
Alongside questions of accuracy, meaning and trust, using artificial intelligence for translation raises a further issue: the protection of personal and sensitive information.
Documents submitted for translation may contain identification numbers, addresses, family details, medical information, financial data or legal information.
When such information is entered into an external technology system, it is important to understand how it is processed, whether it is stored, who may have access to it and what conditions apply to its use.
This does not mean that every artificial intelligence system compromises privacy. There are significant differences between public services, enterprise systems and tools that offer different safeguards and contractual arrangements.
But the mere convenience of a tool does not remove the need to consider whether it is suitable for processing sensitive information.
In this context, a translator’s professional responsibility is not limited to linguistic accuracy. It also includes proper care for the information entrusted to them, its confidentiality and the uses permitted for it.
Is artificial intelligence a threat to translators?
It is easy to understand why the development of artificial intelligence raises questions about the future of the translation profession.
In the past, simply being able to move text from one language to another was a service that required considerable linguistic knowledge. Today, this basic ability is available to the general public with almost no effort.
Technology does not necessarily remove the need for expertise, however. It may change the way that expertise is expressed.
Professional translators have used technological tools in their work for years, including digital dictionaries, terminology databases, translation memories and computer-assisted translation software.
Artificial intelligence can also be a useful aid, depending on the type of document, the level of risk involved and the terms of use of the technology.
It can help find alternative wording, organise information and explore linguistic options. But when a translation is intended for official use, or when an error could have significant consequences, the importance of professional review does not disappear.
In some respects, it actually grows.
The more fluent and convincing the text that automatic systems can produce, the more important it becomes to distinguish between a result that looks right and a result that has been checked and found to be faithful to the source and suited to its context.
ITI’s official position on artificial intelligence also emphasises the continuing importance of the human element in ensuring quality, accuracy and ethics in translation and interpreting services, alongside the opportunities that new technologies offer professionals.
The debate therefore does not have to be framed as human versus machine. A more useful question is how technology can be used without giving up judgement, quality and responsibility.
What really determines the value of professional translation?
The two articles in effect offer two complementary ways of looking at the value of the translation profession.
One focuses on the expertise and linguistic decisions behind a translation. The other focuses on trust, communication and the professional relationship between translator and client.
In both cases, the discussion goes beyond the simple question of how many words can be translated and how quickly.
Professional translation is not just a technical task. It is a service in which decisions are made about the meaning of a document, how it will be read and the possible consequences of its use.
For someone who wants to translate an everyday message or get a general understanding of content in a foreign language, an automatic translation tool may well be enough.
When it comes to an official, legal, academic or important personal document, on the other hand, a higher level of review, expertise and responsibility may well be needed.
This does not mean that every document requires the same process, or that technology has no place in professional translation work. The decision depends on the purpose of the translation, the sensitivity of the information, the requirements of the receiving organisation and the possible consequences of an error.
Being able to tell these situations apart is itself part of professional judgement.
In the end, translation is also a matter of trust
Artificial intelligence will continue to develop, and automatic translation tools are likely to become more accurate, faster and more accessible. There is no reason to ignore the advantages they offer, and no need to present technology as the enemy of the translation profession.
But as the options for translation expand, it becomes more important to understand what service is needed in each case.
Sometimes an automatic translation is enough for a general understanding. In other cases, what is needed is professional translation, careful review, adaptation to the purpose of the document and the ability to answer questions that arise during the work.
A professional translator is measured not only by the ability to find the right word in another language, but also by the ability to understand why that word is right, when another option should be chosen and what might change if the wrong decision is made.
The articles by Lesley Sogl and Dr Christophe Fricker are a reminder that, alongside technological progress, there are elements of the profession that cannot be judged by speed of production alone: understanding of context, responsibility, communication and trust.
Ultimately, when someone hands over a document for translation, they are asking for more than a collection of words in another language. They are asking for the information they have provided to be conveyed correctly, clearly and faithfully to the original.
Technology can help carry the words across. Making sure the meaning has been preserved is the very heart of professional translation.
Sources and references
- Lesley Sogl MITI, ‘AI, Translation and the Value of Judgement’, ITI Insight, October 2026, pp 14-15. Institute of Translation and Interpreting.
- Dr Christophe Fricker, ‘Do Our Clients Trust AI or Us?’, ITI Insight, October 2026, pp 24-25. Institute of Translation and Interpreting.Read the article on the ITI website
- University of Melbourne and KPMG, Trust, Attitudes and Use of Artificial Intelligence: A Global Study 2025.See the study’s findings on the KPMG website
- Institute of Translation and Interpreting, Artificial Intelligence: the Institute’s professional position on the use of artificial intelligence in translation and interpreting.Read the Institute’s position
This article is an independent professional review and analysis inspired by the articles mentioned above. Examples and interpretation not expressly attributed to their authors are part of the independent discussion and are not presented as their views.
Disclaimer
This article was written for information, interest and professional discussion only, and is based in part on publications of the Institute of Translation and Interpreting (ITI), including articles published in ITI Insight magazine.
The article contains the author’s own interpretation, examples, conclusions and opinions, which should not necessarily be taken to reflect the views of the authors of the sources, of ITI or of any other organisation mentioned in it.
All quotations and references to external sources are given with credit to their authors and to the original publications. Copyright in the original content belongs to its owners. Any translations of quotations are provided to make the information accessible and are not official translations on behalf of the authors or publishers.
The article does not constitute legal advice or a specific professional opinion, and should not be relied on as a substitute for a professional review of a particular document, translation or set of circumstances.
It is published independently by Elad Attia Translation Services and is not an official publication of, sponsored by or endorsed by ITI or the authors of the articles.
© 2026 Elad Attia Translation Services. All rights reserved in respect of the original content of this article.