AI for translations – a tempting solution that requires thought
AI for translations – a tempting solution that requires thought
More and more companies are considering using AI for their translations. That is not surprising. The speed is impressive and the price for the translation step itself can drop significantly. But there are also risks with AI translation that are easy to miss, and the steps in the process are many more than just the translation.
In recent years, various AI services have been launched that companies can use to translate texts themselves. However, if you do not have the right expertise, it is better to choose a competent translation partner – this applies to both traditional translation and AI translation.
This post is about what you need to have in place for your AI translations to be cost-effective, work strategically and avoid risks and costly surprises – regardless of whether you translate in-house or via a partner. And it is about what Språkbolaget can offer as a partner in this context.
AI translation requires good source text and careful post-editing
It is easy to focus on the translation process when we talk about AI translation. But the process matters little if the source text is flawed.
A traditional human translator has years of routine, spots flaws in the source text and reports and corrects them. With AI translation, much higher demands are placed on the source text – errors and ambiguities tend to be amplified, and AI does not hesitate to guess if something is not crystal clear. An ambiguous sentence in Swedish becomes even more ambiguous in English, German and French, which means that the time and cost to fix the errors are multiplied.
Therefore, review the source text extra carefully if it is to be machine translated – or let us at Språkbolaget do it.
- Is the terminology consistent?
- Is there an internal glossary, and has it been made available to the AI translation step?
- Is there a style guide – and is it applied?
Naturally, an AI translation must be reviewed by an experienced human translator (known as post-editingOne should be aware that AI services tend to create unexpected errors that are presented with perfect syntax and are therefore very difficult to find – unless you know what you are looking for. We at Språkbolaget know that.
The Institute for Language and Folklore (Isof) states that the quality of the source text is the be-all and end-all for a good translation, and that a terminology list with definitions in the source language is one of the most effective tools you can give a translator, whether human or AI-based.
For a small business, this might involve documenting ten to fifteen key terms. For a larger company with multiple departments, a more structured effort is required, preferably with the support of a terminologist. The investment pays for itself in lower overall costs and fewer review rounds further down the line.
AI translation and data security – what applies?
Many companies underestimate the security aspects when they start using AI in their translation operations, regardless of whether they manage the translation themselves or outsource the task to a language service provider.
The fundamental thing to bear in mind is that AI translation means the text in question is sent to a third party – the AI service.
- Where are the AI service's servers located? Within the EU? In the USA? Somewhere else?
- Is the transfer to and from the AI service's server end-to-end encrypted?
- Is the AI service certified according to the information security standard ISO 27001:2022?
- Is the AI service paid or free of charge? In the latter case, your texts will undoubtedly be used as input for the AI service's LLM (Large Language Model).
Many popular AI services can handle data on servers outside the EU, and with terms that are not always transparent. This can be completely unproblematic or a real business problem, depending on what you upload. A product description is one thing. A quarterly report or an agreement with a strategic partner is something entirely different.
The risk level is one of the most important factors when choosing a translation process. One way to get an overview is to categorise the company's texts according to how sensitive the information is. Most businesses can divide their texts into three levels:
Open information – website copy, product descriptions and certain press releases. Low risk. Often suitable for AI translation.
Intern information – process descriptions, training material, communication with partners. Medium risk. Review the AI service's guarantees and security levels carefully.
Confidential information – contracts, pricing strategies, customer data. High risk. Should not be handled with AI solutions, but by certified suppliers who engage human translators with confidentiality agreements.
The Agency for Digital Government (Digg) is clear on this point: if an employee uses an AI service in their work and processes personal data, the employer may become the data controller, even if no agreement has been concluded with the supplier. The responsibility lies with your organisation, not with the AI service. This applies to organisations of all sizes.
If you are unsure about which documents you can AI-translate – get in touch with us at Språkbolaget. We know our stuff!
How to preserve your brand voice with AI translation
AI translations generates text that is often grammatically correct and fully comprehensible. That is not the same thing as text that sounds like you.
Every brand has a tone of voice, its own way of communicating that makes customers recognise you and trust you. That tone tends to be smoothed out with AI translation. The result is neutral and correct text that works, but which does not speak with your company's voice.
It is clearly noticeable when the context changes. For a local company, it can mean that the personal tone that makes you unique in your domestic market is lost when you start communicating in English. For a national company, it is about maintaining a consistent language internally and externally even as volumes increase. For an international company, the challenge is to ensure that the tone of voice is consistent and recognisable across all markets, rather than simply having the text translated into the respective languages.
These are aspects that even advanced AI tools cannot handle. This requires human translation – or AI-assisted translation with stringent human post-editing, consistency checking and quality assurance.
Professional translators use their own or clients' databases (known as translation memories) to reuse approved text, and term bases to ensure your internal terminology is maintained. Professional language specialists make sure the end result actually sounds like you. It is the combination of the speed of AI services and human expertise that delivers volume and quality. Finding the right balance, the most effective collaboration between human and AI, and the best value for money requires knowledge and experience – and that is what Språkbolaget can contribute.
Three questions before you choose AI translation
Before deciding on an AI translation process, it is worth pausing for a moment. These three questions usually quickly reveal which solution actually suits the business.
- How does the text look today – is it completely finished and without any doubts?
- Which of the organisation's texts contain information that must not be distributed?
- How important is it that the organisation's voice and tone live on in the translated text?
The answers to these questions often govern which translation strategy you should choose.
Discuss your company's translation strategy with us!
At Språkbolaget we combine advanced translation technology with human expertise and many years of experience. We help assess what is suitable for your texts, your industry and your data security requirements, whether you are a small company taking your first step into a new market or an established business looking to streamline an existing workflow.