Technology to support writing

The most common office software and collaboration platforms (e.g. Microsoft Office 365) include a translation tool that can be used to translate a document or message into a chosen language or to help you write texts in another language after first writing it in your strongest language. The ability to use translations speeds up communication: on one hand they enable access to content in your strongest language and, on the other, they help you to check whether you understood a foreign-language text, thus supporting language learning. However, machine translations are often imperfect, especially in situations that require cultural knowledge and contextual awareness. In fact, all translation supported by artificial intelligence should only be seen as an auxiliary tool, or “support intelligence”.
 

Microsoft 365 software, such as PowerPoint and Word, has come with an AI-assisted translator for years. You can translate an entire document, word, or selected text area at a time. The same artificial intelligence assistance is also available for, among others, Outlook and Teams.

The programs also include a proofreading tool that helps check the spelling and grammar of the written text when the language settings of the document are set correctly. The more important the document, the more important it is to check the final result yourself. However, automatic translation and proofreading programmes can support language learning and working in another language, such as by offering alternative phrasings or concepts.
 

By clicking on a word marked with  red, the user can choose what the AI thinks is a better spelling or ask to ignore the word. You can also add a word to your personal dictionary, which remembers the new word and no longer notes the spelling.

Data privacy and content classes
 

Be mindful of data privacy when using online machine translators. Many online machine translators store all processed data. It is therefore important to start by defining which content class the translated text belongs to. For example, Finnish higher education institutions have introduced the following classification:

1. Public content (e.g. something found on a website)
2. Internal content (e.g. instructions that can be found on the organisation’s intranet)
3. Confidential information, with or without personal data (e.g. email)
4. Secret information (e.g. medical records)

Web-based, unprotected machine translators may only be used when translating public content. Organisations may also use a machine translation service based on their own language model, such as ChatGPT. In this case the service is trained to handle texts specific to that organisation. These internal services are often safe to use for translating work documents. For example, the Aalto AI Assistant developed at Aalto University is based on OpenAI's GPT language models, but it has been incorporated into Aalto's systems in a technically and contractually protected manner, which allows handling not only public content but also internal and confidential data. Correspondingly, the staff of the University of Helsinki have a secure translation tool, Kontra, for translating between English, Finnish and Swedish. 

Author tools

In addition to Google and office software, machine translations include

• eTranslation by the European Commission (registration required)
• DeepL Translator
• Machine translator by Fiskmö project (translations between Finnish and Swedish/Norwegian/Danish)
• MOT Kääntäjä. 

Grammar and spell-checking tools include

• DeepL Write
• MOT Kielentarkistin.

Dictionaries and term and translation banks

When writing or reading in your weaker language, you can also use dictionaries as well as term and translation banks found online (payable or free of charge). Here are a few examples:

• MOT Kielipalvelu: dictionaries for different languages and fields, subject to fee
• TEPA term bank (TSK): terms and their definitions for special fields of interest, mostly in Finnish, Swedish, English and German, free of charge
• The Helsinki Term Bank for the Arts and Sciences: terms and their definitions, equivalents in different languages for different scientific and scholarly fields, free of charge
• Finlex: Finnish legislation translated to different languages, free of charge
• IATE (Interactive Terminology for Europe): EU terminology in different languages, free of charge.

Technology assistance for spoken situations

Following oral presentations, such as Teams meetings or other presentations can be difficult to follow without full language proficiency. On the other hand, the presentation usually benefits from the speaker using their strongest language. Technology can help streamline these situations.

Real-time captions provide an instant transcription of the spoken content in the same language. This feature is included in Teams and PowerPoint, among others. Seeing speech as text often benefits especially non-native speakers.

Translated captions allow listeners to follow presentations by reading captions in the language of their choice. The speaker may also choose the captioning language beforehand. This feature is included e.g. in PowerPoint Live in Teams. 

Some software, such as Panopto, also allow their captions or translated captions to be edited afterwards in case the recording is used later. Note: According to accessibility guidelines, recordings intended for permanent use must include captions in the source language (in other words, a recording in Finnish requires Finnish captions).

Remote meeting platforms (e.g. Teams and Zoom) often also include a simultaneous interpretation feature so that listeners can choose whether to listen to the original presentation or an interpreter.

Meeting transcriptions provide a written recording of the speech. They can also identify speakers and timestamp each part of the discussion. This feature is included, for example, in Teams.

While these AI-based tools provide considerable support, it is important to note that they are not perfect. Users must be aware of potential errors in translations and inaccurate captions. Captioning and translation quality can be improved by using high-quality microphones and audio devices as well as a strong Internet connection.