January 2019
Text by Yasmin Steuert

Image: © SIphotography/istockphoto.com

Yasin Steiert works as an English editor with Afaf Translations. He graduated from the American University of Sharjah, UAE, with a degree in International Relations in 2016. He is fluent in both Arabic and English.



Facing Artificial Intelligence as a small language service provider

The rise of Artificial Intelligence (AI) is threatening many freelance translators and small language service providers (LSPs). But worrying about declining prices is no solution. What is needed is a collective approach to protect small businesses and to use AI to secure their existence.

Artificial Intelligence (AI) is the buzzword you’ve probably heard enough of this year. General technological forecasting paints an ugly picture of an invaded labor space where man and machine coexist in an inharmonic fashion while businesses focus on hardware efficiency. At least, this is the regular journalistic conflation that is based on a phenomenon known as the "panic cycle". Figure 1 depicts this cyclical reactionary trend that has been witnessed in the face of new technologies.

Figure 1: The technology panic cycle
Source: Information Technology & Innovation Foundation


This is not to say that concern and panic are unfounded. It is an economic fact that not all parties will benefit from new technological developments. In the U.S., factories have already been heavily impacted. This, along with indications of poor political economic management, has led to an increase in layoffs.

Repetitive and rule-based labor roles are fast approaching their lowest point. This perception is so strong that some believe that Michigan, Wisconsin, and Pennsylvania would have swung towards electing Hillary Clinton if automation adoption had been two percent less prevalent in these states between 2012 and 2016. All around the globe, we have witnessed ruthless politicians stoking fears to take advantage of existing anxieties. These economic concerns are even more present in the language services industry. In an industry, where all input and output depend on a dual relationship, AI can naturally have a far deeper impact.

In its forecast for 2018, the language service industry research firm Common Sense Advisory highlighted the issues of stratification and consolidation, which are occurring globally as multinational providers seek to scale up in order to meet declining price demands. As human resources in the translation market continue to be replaced by automated systems, it is vital for freelancers and smaller LSPs to learn what policies work, and to know the institutions that may serve as a platform for their voices.

Protecting freelancers in a global market

Over the past century, labor protection movements have seen adequate levels of convergence across different sectors at different periods. Yet, what is often neglected when talking about labor rights is what motivation lies behind this striving for convergence. The incentive to be part of the globalized market and to participate in it with prosperous economic powerhouses remains a primary motivation because, to do so, governments typically need to reinforce labor compliance or else remain vulnerable to consumer backlash and domestic instability.

The axis of Anglo-American and European values has provided two labor environments that will likely constitute the precedent for global freelancer protection. Looking at current trends, it appears the European market will be at an advantage, but just as migration to the "New World" increased wages in Europe and decreased them in the continental U.S., the same might occur on a global scale. Any nation that leads the push towards freelance labor protection will result in language service providers probing periphery nations for lower-cost, financially exploitable, unprotected labor.

Anecdotally, the market seems to be in consensus that due to the strong competition, the overall decline in pricing has become so fierce that award decisions are created on quotation differences that vary by a tenth of a cent.

On a microeconomic scale, local governments have already shown initiatives to continue to provide market power to smaller actors on the LSP market. When going through the contract award process with the city of San Francisco, it becomes clear that the city government advocates for small LSPs that can no longer compete adequately in the national market. The San Francisco Local Business Enterprise (LBE) makes sure that in the face of large companies, local businesses have a shot at the competition. In the current environment, the LBE removes the expensive barrier of entry that AI entails.

Is there room for fair-trade translation?

There is no moral justification for companies leveraging regions under economic and political oppression to find the lone language service contractor willing to accept the work because it is his only means of income. Unfortunately, there is little interest in the welfare of someone who is only known through an occasional email and is located thousands of miles away.

There is no fair-trade initiative in language services, and it is clear that the diverse, multidisciplinary foundation that made this industry flourish into what it is today is ultimately being consolidated without considering the impact. Another taxi driver without a car, another farmer without land, and another translator without words. This displacement is detrimental, both psychologically to the individual as well as to the communities. The aforementioned panic cycles further aggravate this situation by distorting the actual threat of AI without providing helpful solutions. It is clear that small to mid-sized LSPs are happy to accept cheaper rates for a good number of their highest volume languages. Margins are shrinking and the advantage of overhead costs will no longer matter for the smallest actors if AI software remains protected.

Unfortunately, apart from initiatives by local governments, protective developments for freelancers are lacking entirely. Common Sense Advisory recently named TRAVOD as the largest LSP in Europe. TRAVOD translation rates are around .035 Euro per word as an average starting price and only go down from here. There is no doubt that, as a company, TRAVOD is doing fine with its management and price target initiatives; that’s what makes them the largest LSP in Europe. But this process of scaling is not merely the result of "technophobia", as certain AI commentators put it.

What makes a healthy global economy?

Scaling with technology that is not accessible to the average professional entity is going to result in consolidation. We’ve seen this as a common trend across almost all sectors in the U.S. And though this consolidation has been accompanied with increasing employment, average wages are at their lowest. Yes, on paper the global economy is healthy, but its environment is increasingly noncompetitive and unequal. Protections against corporate consolidation are failing and though the market is "healthy", consolidation will end this. The European Trade Union Confederation has already raised the issue of freelance worker protection, which extends from the typical freelancer to those "companies" that are merely sole proprietorships. This is a step in the right direction. Overall, it is more common for local governments in Europe to employ contract award practices that are similar to those of San Francisco’s LBE. This is what ultimately should be done on a global scale. Addressing AI and understanding the changes to our modes of production is our first challenge.

The future of translation

It is necessary to talk about re-education programs that need to be dispensed in the LSP community. LSPs that are not capable of competing by scaling will have to direct investment towards shifting labor tasks and responsibilities. It is very likely that the future LSP environment will be more decentralized, and AI could act as a complementary tool in any translator’s toolbox, allowing the independent worker to generate an output that equals that of major companies today.

But in order to attain such an environment, small and medium-sized LSPs need to consolidate their own power. The top LSPs will most likely not share their wealth, as they are trying to keep their edge over the competition, increase their market share and ultimately control the market price. LSPs need to find fellow translators from other fields, sole proprietorships and freelancers from all industries, and each individual must stand up in arms against the massive gig economy that will continue to grow. LSPs that are worried about the marginalization of project availability and declining prices will have no voice as constituents alone. But together, we have bargaining power, and together, we can overcome the precarious side effects of new technologies and use AI to secure our financial existence and collective wealth.