The market for interpreting services has gone through some turbulent developments over the last decade. We’ve seen the rise of new technologies that have threatened the existence of traditional interpreting services. And while there’s been a lot of work on that front, it still seems like the current state of technology is far from advanced enough to provide a full replacement.
The Future of Interpreting Services
Major Developments Over the Last Decade
The last decade saw the rise of artificial intelligence and technological assistance in general. Solutions like machine learning have allowed many industries to make significant advances in their work, including translation, interpretation, and writing in general.
Interpreting Services
GPT-3 was a recent major milestone in text generation in particular, touted as having the ability to produce text that’s indistinguishable from something written by a human. DeepL started out as a minor competitor to Google Translate. But has been progressing at a rapid pace. And Google Translate itself has benefited a lot from the integration of AI-based technology. Particularly in the way of translating between arbitrary language pairs that don’t include English.
Why Are Human Interpreters Still So Popular?
And yet, despite all that, human translators and interpreters have retained a leading position in the market and have not shown any sign of slowing down anytime soon. There was a lot of talk about interpreting services getting eventually overtaken by AI-driven solutions. And yet we’re still far from that situation.
Translators and interpreters
This can be attributed to multiple factors. Some of them are major bottlenecks in the growth of AI-based translation and interpretation, and it looks like those challenges will not disappear from the horizon anytime soon.
AI Is Still Far from Perfect
Modern AI-based translation and text generation are very impressive at a glance. Just throw a small prompt at an AI system and it can produce an article of any length that seemingly goes into a lot of detail on the subject it’s been provided. Solutions like these have also been integrated into modern translation engines, increasing not only the base accuracy of their translations, but also the readability of the final text.
AI Is Still Far from Perfect
And yet, that magic wears off pretty fast. Once you’ve worked with an AI-driven system for just a little while, you’ll immediately start to spot some patterns in the way it operates. These issues get particularly obvious once you start testing the system on more complex input and not just simple articles that can be produced by rehashing the top 10 Google results on a given topic.
The sentence structure and grammar may often be close to perfection. But the underlying flow and delivery are still very far from that point. It may work as a minor trick to impress one-off clients. But long-term clients will quickly catch on to what an agency is doing if it primarily relies on AI for its translation and interpretation.
Access to Cutting-Edge Tech is Restricted
The best solutions on the market are not only far from perfect, they are also very restricted in terms of access. GPT-3 is only available through an Open AI subscription. It starts out relatively affordable. But can quickly end up very expensive when used commercially. It’s also possible to download, train, and run the model on your own, but this requires hardware that costs tens. If not hundreds of thousands of dollars.
Access to Cutting-Edge Tech is Restricted
Things will likely stay that way for a long time. Even if a better solution comes out in the near future, it will likely be even more restricted than GPT-3 is at the moment. That’s partly because the research teams behind these products see themselves as having a certain moral obligation to prevent their tools from becoming publicly accessible to avoid having auto-generated fake news spreading like wildfire.
Cultural Subtleties
AI is good at understanding the basic intention of a piece of text. But it’s still very far from the ability to grasp the more delicate subtleties of the languages it works with. Simple puns often make it through the translation successfully. But if you take a whole paragraph with a more metaphorical meaning, things quickly start to fall apart and the cracks in the foundation become obvious.
Translators and interpreters
Successful interpretation relies heavily on the ability to understand these subtle points of the languages involved. And to translate them as well as possible into the target language. Judging by the way things are right now. It will take a long time until AI is capable of translating complex pieces in a way that preserves every detail of the original.
The Issue with Dialects and Accents
And that’s just the written part. When it comes to speech – another integral part of interpretation services – AI gets even worse. AI is still terrible at speaking with a proper dialect and accent where it would make sense to enrich the original interpretation.
Interpretation services
Many people are already regularly expressing annoyance at the robotic female voice associated with videos published on a certain well-known social media platform because of how repetitive it sounds after a while. And that’s supposed to be a cutting-edge solution at the moment.
Human Pronunciation Is Still Important
Apart from dialects and accents, humans also have various small imperfections about their delivery. Which remain hard for AI to mimic. Having your interpreting done by an actual human is obvious right from the start. And those small details that we often throw in unconsciously (or, in the case of an experienced interpreter, sometimes very consciously) can take the customer’s impression to the next level and really drive home the point of how well the interpreter is doing their job.
Dialects and accents
Will this be successfully replicated in the future? Possibly, but likely not anytime soon. And when it does, we will be brought back to the issues we described above, including availability, pricing, and other obstacles in the way of giving this technology an established presence on the market at large.
Will Interpreting Services Ever Get Fully Taken Over by AI?
In the end, artificial intelligence remains a major potential competitor to traditional translation and interpreting services. However, we’re still far off from the time when this will be a full reality. It will likely take at least a decade from the current point to get there. And that’s if the industry doesn’t encounter any unexpected obstacles that set it back.
Artificial intelligence
Until then, human interpreters will remain the primary choice for those who insist on receiving a high-end service that covers all their requirements as well as possible.
Final Thoughts – Interpreting Services
Artificial intelligence is not to be underestimated, and the threat it poses to the interpretation market is very real. But as anyone who’s used this type of service extensively can readily tell you. It’s very far from being able to provide a reasonable substitute for human interpreters. In addition, certain implementations of AI tech in interpretation provide great assistance tools to human interpreters. Further improving their ability to work on complex tasks. It’s very likely that this will be the primary use of AI in interpretation for a long time.