Are you assessing, measuring, and mitigating language risk within your organization? AI, chatbots, machine translation, and other forms of automation make it super easy to translate from anywhere, with varying levels of fit for purpose. AI is not magic. The small print (and indemnification clauses that your lawyer should have checked) for all LLMs carries disclaimers about accuracy and shifts responsibility for any errors to the user.  

Some content types require a linguist to sign off, or certify, the accuracy of content transformed from one language to another. Many language pairs simply have insufficient data to enable any MT engine or LLM to be trained in a meaningful way. 

On the other hand, well written and consistent content for which errors carry little risk of causing harm can prove highly successful in fully automated translation processes, especially for high-resource languages and those with well-defined terminology and style guidelines. 

You can’t just generalize and put all of your content through the same process, assuming the same level of accuracy, and therefore risk, for every type of content and every language pair. 

The Issue: Translation Decisions Require Risk Assessment 

To make informed decisions, you need to conduct a thorough content audit that includes a risk assessment and a determination of which risks are acceptable to the company brand, customers, employees, and regulated status. 

How Language Risk was Assessed Before AI 

I remember sitting down with a top-level compliance manager at a major high-tech company, many years ago, before MT became mainstream. The term “post-editing” was barely known then. AI was science fiction. 

We met to discuss translation of user documentation. I was there for localization strategy; he was there to assess risk to the company. I’ll call him Mr. Compliance. 

We went through all the regulatory requirements my team had uncovered for the delivery of products and documentation in local languages for the countries in which the company did business. Not all markets had such laws or regulations: some were detailed, some were vague, some were implied, and some were non-existent. We also knew the cost of translating and producing localized versions – some of which involved serious libraries of heavy, printed, loose-leaf manuals which had to be physically shipped – and then continuously updated with each software patch or release. 

For each country, we examined the requirements, and then Mr. Compliance assessed the risk: 

And so the conversation continued. 

The focus was on cost, legal compliance, and likelihood of enforcement. In many cases, the cost of localization far outweighed the potential penalties. User experience never entered the equation. 

What has Changed: Language Risk is Broader Today 

You still need to run this kind of assessment today, but with additional inputs and considerations based on the potential low cost of variable-quality translation, investment in human expert translation and validation, and a deeper understanding of what language means in communication with your audiences and customers (“Can’t Read, Won’t Buy – B2C” and “Can't Read, Won't Buy - The B2B Edition”).  

Regulated industries still face increasing, more complex, and stringent checks for accuracy. At the same time, AI introduces low cost but variable quality translation options. 

Language risk management is no longer simply about compliance with a small set of import regulations. It is also about: 

Unlike the earlier conversation with Mr. Compliance, it’s no longer a binary decision based on the high cost of translation versus the likelihood of enforcement and the cost of a penalty. Enterprises now have a more nuanced set of options and can deliver in-language content even where the cost of implementation previously prevented it. You can be more compliant, without breaking the bank – if, and only if, you apply language risk best practices. 

Implications: Match Language Risk to the Translation Approach 

The key question is whether you understand your language risk landscape. 

Do you know which of your content types, including target audiences, regulatory requirements, and language pairs are high, moderate, or low risk? 

Have you matched these risk profiles to the right approach? 

Language risk assessment should guide how you route content to AI, human, and hybrid translation workflows. A structured approach to language risk allows you to balance cost, quality, compliance, and customer experience, and to make better decisions about how to translate your content. 

This series of reports may help you take the right steps: Everything You Need to Know About Language Risk Management