Mission

To keep translators and other language professionals embedded in the highest quality translation workflows in the industry

Goals

  • Provide a tool that lets translators and other language professionals do more with AI than they would with other solutions alone
  • Train translators and other language professionals to be more productive with AI integrated into their workflows
  • Foster a strong translator and language professional-based center within the localization industry

Our Team

Steven S. Bammel, PhD

Steven S. Bammel, PhD

Steven is a Korean-to-English technical translator with over 20 years of experience. Having a background in Strategic Management from Hanyang University (Korea) and in Economics from the University of Texas at Arlington (USA), Steven combines his language and translation expertise with a deep knowledge of AI workflows to develop frameworks that help translators raise their productivity. As an active member of the American Translators Association (ATA), Steven still translates daily.

Main work location: Hongseong, Korea

Stanislav Okhvat

Stanislav Okhvat is a translator turned software developer with a background in language studies and technical translation. During his career as an in-house technical translator, he worked to increase the productivity of the company’s translation team. This led him to develop TransTools, a suite of productivity tools for translators and editors, and TransTools+, a set of additional advanced tools for translation industry professionals.

Main work location: Tbilisi, Georgia

Dave Neve (B.A. in Linguistics)

Dave graduated in Linguistics with Swedish and French at Newcastle University (UK). He went on to teach English as a foreign language in Spain and then France, mostly in higher education. Having enjoyed translating for colleagues, students and educational institutions, Dave became a full-time freelance translator in 2013, under the name “SafeTex”.

Dave joined the CotranslatorAI team in July 2023 as a voluntary consultant, and now focuses on non-translation market promotion and community development, while continuing to translate part-time.

Main work location: Lisle, France

Our Story

Entering the translation industry

Steven first learned about the world of translation from another translator back in 2000. The industry was full of new technology, like CAT tools, email, and the Internet. These brought exciting global opportunities.

Becoming a translator changed his life. Steven was motivated by the need to support his family and have a steady career. He was also excited by the new and growing market he saw.

It was an exciting time to be a translator…

Or at least, that’s what it was on the good days.

Even though he loved working with languages, the ups and downs of being a freelance translator were difficult. To deal with this, Steven worked hard to improve and become more competitive.

He enhanced my knowledge in specific areas, got advanced degrees and translation certifications, learned new workflows, bought new software, and tried offering services like revising and copywriting. He also marketed himself online and offline, to both direct clients and LSPs. He tried everything.

His goal was to increase his earnings and have more steady income.

Have you been in this business long enough to understand this struggle?

These efforts helped to some extent and for a while.

Navigating technology and its limits

But he discovered that trying to keep up with new technology was held back by an economic reality. As technology got better, the extra returns from those innovations didn’t reach translators.

Still, around 2020, Steven had a big realization!

He finally saw that machine translation not only improved his efficiency (at least sometimes), but he could use the machine to deliver better work!

This doesn’t mean the machine translation was more accurate or better written than his own translations.

But sometimes it produced a perfect result (which saved time), and even when it didn’t, it gave ideas that added variety to his writing.

Steven researched and found a so-called “adaptive machine translation” plugin for his CAT tool. They claimed it learned from previous data and provided MT that was more consistent with his past translations.

Steven thought that was fantastic!

It didn’t always work as promised, honestly. But occasionally it did, and he was still getting that “variety” boost just mentioned above.

Steven translated with that plugin for a couple of years, fairly satisfied.

The Generative AI breakthrough

But in late 2022 when he tried translation with the new ChatGPT chatbot for the first time, it amazed him!

Why? Because it translated perfectly?

Nope.

Because it wrote more naturally than the adaptive machine translation plugin?

Yes, partly. But that wasn’t what was most impressive.

What amazed Steven was that Generative AI has superpowers, and if he managed the process correctly, he could almost eliminate internal inconsistency and produce translations in his own voice, not the machine’s voice!

Within a month, he canceled his adaptive machine translation plugin subscription and fully switched to Generative AI.

Generative AI is much more powerful than just as a replacement for primitive machine translation; we can apply it across all of our roles as translators.

Because of this, it became clear to Steven that this new Generative AI was a game changer in our profession.

He also saw one more key thing….

He saw that Generative AI is a democratizing technology.

While we had always relied on our clients and industry innovators to pass their solutions on to us, with Generative AI, we have access to the power directly and affordably, able to mold it to our purposes.

Steven knew he had to figure this out, and so he began a journey of discovery.

Overcoming AI’s shortcomings

As Steven investigated Generative AI, he saw how it hallucinates and mistranslates, that it often writes like a machine, and that the ChatGPT privacy policy is weak. Cutting and pasting text back and forth between the ChatGPT interface and the work environment was also inefficient.

At the same time, he developed a rudimentary technique to transcend the segment-by-segment paradigm of the CAT tool while still working with the CAT tool. He got ChatGPT to follow instructions on terminology and style, and improve with a few example segments. Steven discovered if he kept improving his prompt, the translations he got from ChatGPT got better, and that when everything came together, it started to translate in his own voice, not in the voice of the machine.

This was the epiphany. Steven had to find a way to bridge the gap between this incredible potential and all the shortcomings of Generative AI that were becoming more evident by the day.

Building a better solution for translators

Steven discussed his predicament with a colleague, Stanislav Okhvat, a programmer who has developed leading software for translators. Stanislav had an idea…

He said he could build a simple tool inspired by Steven’s methodology that would overcome all of these shortcomings of Generative AI.

First, he explained that this tool could solve data privacy concerns by going through the OpenAI API, rather than ChatGPT. With the API, all data from the user’s computer would travel to the OpenAI servers through an encrypted channel and that no client data would be used for training, and that the data would be deleted from the OpenAI servers promptly.

Second, he said that this tool could leverage the context window in Generative AI with a unique interface design that would crack the dominant segment-by-segment translation paradigm in the industry. Steven wasn’t sure exactly what meant or how it was going to help at the time, but it sounded cool.

Third, Stanislav suggested adding a prompt library to the tool where translators could write and store complex prompts customized to every use case they needed Generative AI to help with. Stanislav explained that translators could iteratively improve these prompts, effectively storing their tacit knowledge in a ready-to-use format.

Finally, he devised an ingenious system of keyboard shortcuts to allow a translator to apply any of their prompts to any selectable text in Windows with just a simple key press.

This was the purest channel into the AI anywhere, and they knew now that they weren’t just building a better tool for machine translation; Generative AI in CotranslatorAI would transform every task of the translator.