(Updates with additional comment from Lilt CEO in paragraph 8.)
Bloomberg Government subscribers get the stories like this first. Act now and gain unlimited access to everything you need to grow your opportunities. Learn more.
Training manuals and tutorials for the weapons heading from the US to Ukraine are piling up, with few Americans able to translate them into either Ukrainian or Russian.
There is a limited pool of American citizens who speak Ukrainian and who can get a top-secret clearance to process national security-related documents. US trainers have acknowledged some of the challenges in teaching Ukrainian troops to use the millions of dollars of equipment provided by the US and other allies.
“There are the natural challenges associated with training in a second language, everything from overcoming a language barrier to some of the technical piece,” said Brig. Gen. Joseph Hilbert, head of the 7th Army Training Command in Europe, who is overseeing the training of small groups of Ukrainians in Grafenwoehr, Germany.
But the federal government spends nearly 200 times more on procurement for translation and and other language requirements that rely on people, than it does on AI-based services.
Artificial intelligence language translation company Lilt wants to solve the government’s language gap with its technology, said CEO and founder Spence Green.
The US government has 83% of required staffing levels for Slavic linguists, and there is a need for at least 100 more Slavic linguists just for the US intelligence community alone, according to Green’s conversations with US government officials.
“You’ve got all this information that suddenly has to go into Ukrainian, and you have some fixed pool of people in the United States in a classified environment,” Green said. “One solution is to train more of them. The other solution is to use machine learning to make that small group of people five times as many people.”
Translating on the Battlefield
“The solution of sending people to language school for 12 months does not solve President Zelenskiy’s problems,” Green added, referring to Ukrainian President Volodymyr Zelenskiy.
The Defense Language Institute Foreign Language Center in Monterey, Calif., and Washington, D.C., conducts immersion courses for land military intelligence jobs and positions within the FBI and the National Security Agency. The languages taught depend on the needs of the services and the Department of Defense, according to Natela Cutter, spokesperson for the institute.
The center currently doesn’t teach any Ukrainian classes, with high enrollment classes in Chinese, Korean, Russian, Arabic, Spanish, and Persian Farsi, she said.
“I don’t think one could say that there is a ‘gap’ in Slavic languages,” Cutter said. “It is fair to say that after the fall of the Berlin Wall and the changes in Eastern Europe in the early 1990s, we stopped teaching a number of East European languages.”
Rusty Shughart, a retired Air Force colonel who serves as the chair of the American Translators Association’s Government Linguist Outreach Task Force, said handling language crises can be difficult in scope and duration, and must involve not just the government and the military, but also volunteer language specialists, US allies, and non-governmental organizations, among others.
But he was skeptical that AI could work in Ukraine.
“AI is making some pretty significant advances, and especially in machine translation,” he said. “I am very impressed with what AI can do, especially with other related areas like neural machine translations and so forth, but I do not think we are at the point where we can apply these things to this particular battlefield.”
Market for AI Translation
Since fiscal 2018, the contracting segment of the Bloomberg Government-defined artificial intelligence and machine learning market for translating has amounted to less than $25 million, including $2.3 million to Lilt. The total for all translation and lawful intercept services in that time is $4.7 billion.
Competition in the US startup market for translation services is slim, with most of the space dominated by big tech companies like Microsoft and Google. Green pointed to a Chinese company with $400 million in state funding as one example of non-US development in AI-based translation. Baidu Inc., a Chinese company that hosts one of China’s largest search engines and other AI projects, is developing a portable translator.
Lilt’s pitch is that it keeps “humans in the loop” of the AI algorithm to translate written content—a move that allows the company to provide translation at a lower cost than hiring human translators, and more accurately than using just AI.
Interview requests to AFWERX, a technology directorate of the US Air Force that awarded Lilt with a Small Business Innovation Research contract, weren’t immediately returned.
But Intel Corp. invested in Lilt in 2019 and they’ve been working with them since. Mark Rostick, vice president and senior managing director at Intel Capital, said he saw Lilt as a “really exciting new company that was really trying to change a market that hadn’t changed in 30 or 40 years.”
Rostick described the larger translation platforms by big tech companies as “in the good enough phase,” but added that Lilt’s ability to learn from a specific data set and its level of accuracy makes it a challenger in this space.
“There are times you can’t afford to make a mistake,” he said regarding accurate translations.
Lilt’s human-in-the-loop technology can translate more than 70 languages and currently contracts with the intelligence-sharing alliance known as “Five Eyes,” made up of the US, Canada, the UK, Australia, and New Zealand. It also partners with the US Air Force, Navy, and the FBI.
The language agnostic and data-oriented system learns from “parallel texts”—meaning processing a sentence in English and the same sentence in Ukrainian, for example. The software is self-learning, and doesn’t require a network connection, which distinguishes it from some of the other major players in the space.
Lilt is currently working with Invariant, a government affairs and public relations firm run by Heather Podesta, to get more government support, including in the Ukraine emergency funding request. The $40 billion Ukraine supplemental the House voted on Tuesday didn’t set aside any specific funds for translation, but mentioned broader reimbursement for Pentagon “defense services.”
The company is also focused on fiscal 2023 funding for classified programs, as well as research and development accounts that may be part of the annual defense authorization and appropriations bills Congress is working on.
Lilt has already spent $130,000 in lobbying fees at Invariant since late last year, according to the Senate lobbying disclosure database. Lilt previously spent $320,000 with Banner Public Affairs LLC, according to the database.
Lilt and Invariant are meeting with lawmakers and government officials “to make the point that with an overwhelming flood of foreign language content and data in the digital age, national security agencies simply cannot hire enough qualified, cleared American translators to brute force their way out of the problem,” Green said in an email.