AI-First Interpreting Approach: What it is and Why it Matters

The field of multilingual communication is undergoing a profound transformation. This transformation is rapidly ushering in an AI-first paradigm, a reality in which AI will be the default agent for translation and interpreting services. This shift is not distant or speculative; it is already underway, albeit still in the early stages for real-time spoken translation. Yet, many institutions and policymakers remain unaware of or unprepared to address the scale and implications of this change. The same can be said for stakeholder groups that relied on a human-first paradigm, such as interpreters, interpreting scholars, and representatives of the professions. Many of these groups have yet to fully grasp or address the disruptive potential of AI-driven interpreting technologies.

What Is an AI-First Interpreting Approach

Across public institutions, procurement agencies, and even regulatory bodies, there remains a deep-rooted assumption that interpreting services will remain predominantly human-delivered. This view is echoed by interpreting academia, professional associations, and practitioner communities, many of whom profess the belief that interpreting, due to its human complexity, should and will remain untouched by automation. 

While this assumption is understandable, it is becoming increasingly misaligned with the current trajectory of technology and the market. AI interpreting tools are already in use, often without proper governance, oversight, or established standards. For now, human professionals are still preferred, while AI is generally regarded as a supplementary or experimental option that is still somewhat distant from mainstream practical use. This is because the quality is undoubtedly subpar compared to professional services.

However, a plausible scenario is emerging for the near future: one in which machines perform at a level that resembles human interpreting. In this context, the market will quickly shift toward an AI-first model where automated tools are the standard solution for overcoming language barriers. Human professionals would then primarily be called upon in cases where AI fails to meet expectations.

This kind of shift is very common in history. Consider travel planning. Decades ago, booking a trip required a visit to a human travel agent. Now, online booking portals like Expedia or Booking.com have become the default, offering instant access to flights, hotels, and rental cars. Travel agencies still operate, but they typically serve specialized markets such as luxury travel or complex itineraries.

Technology can redefine an entire service model: the traditional version doesn't vanish entirely, but becomes a specialized option while automation leads the mainstream.

Why an AI-First Interpreting Approach is a Plausible Scenario

With the quality of real-time speech translation improving and forecasted to increase in the next few years, AI is set to become increasingly appealing to many stakeholders in need of interpreting services. It is my conservative prediction that AI interpreting will become virtually indistinguishable from human interpreting by 2030. Besides quality, multiple non-technological factors are pushing interpreting toward an AI-first model. They are economic and practical:

•    Cost reduction pressures: Clients want faster, cheaper access to multilingual communication. Human interpreting, with its scheduling complexity and high fees, cannot scale affordably.
•    AI as a productivity multiplier: In broader narratives, AI is framed as a force that increases output while decreasing cost. Translation and interpreting are being rebranded as processes that can (and should) be optimized.
•    Frictionless deployment: AI interpreting can be integrated with a click. No scheduling, no travel, no availability issues. For many users, especially in the private sector, that alone makes AI more attractive.
•    Lack of interpreters: In some sectors, public institutions express lack interpreters to cover their often-increasing needs. AI is seen as a viable alternative to meet this challenge. 

The result will likely be an increased demand for AI-mediated interpreting services. This demand will not only arise in settings that were previously untouched by interpreters, as is mostly the case today, but also in settings that were previously served by human professionals exclusively.

This transformation is already changing, albeit insignificantly for the time being, how interpreting is consumed and provided. Casual tools like Google Translate are used to overcome language barriers when no other solutions are provided, for example, by police or hospitals. Specialized, on-demand, AI-powered interpreting services are being integrated into platforms and tools where users may not even realize AI is at work. The forecast is that human interpreting will become a premium or niche service reserved for settings where high trust, legal stakes, or AI-unsupported languages still matter deeply.

Risks of an AI-First Interpreting Approach: When the Gap Narrows, the Stakes Rise

As we have argued, AI systems are rapidly improving and will increasingly match or exceed human performance in many interpreting tasks, at least in terms of outward appearance. The goal, whether stated or implied, is indistinguishability from human interpreters: passing what I called the Turing test for speech translation. As with the many areas where generative AI is becoming indistinguishable from reality (e.g., deepfakes), a similar outcome will be reached in interpreting. When this occurs, the temptation to use AI translation systems indiscriminately will be high. This could lead to issues since stakeholders are not prepared to govern this use.

Although quality will improve to a level that inspires confidence, subtle differences will persist. In some contexts, those differences matter. For example, in high-trust or emotionally charged situations such as medical consultations, asylum hearings, and corporate negotiations, AI may lack the ability to perform the broader mediation work that human interpreters naturally do. Sometimes, translation alone is insufficient; an active mediator is needed. While it is conceivable that advanced AI will one day achieve this, it is a different matter entirely from achieving high-quality translation. For low-resource languages, where AI training data is scarce, AI interpreters will be of poor quality. Ironically, while Meta and others work toward "no language left behind," we risk abandoning these languages to the margins. Human interpreters will remain essential here, but the challenge will be informing policymakers of these differences and, more importantly, objectively measuring the quality of translation systems to make informed decisions.

At the moment we are not ready for this, and the risks is simply an indiscriminate use of AI systems. Without robust quality assurance, deploying AI in sensitive contexts could cause serious harm. Errors may go undetected (as it is the case for humans), and the illusion of perfection may breed dangerous complacency. In regulated markets, policymakers must advocate for the regulated use of such systems. To do so requires expertise.

Conclusions

AI-first interpreting is no longer a distant possibility. It is a plausible, even probable, near-term future scenario. Pretending otherwise is no longer tenable.

Policymakers must do more than indulge the temptation to experiment with AI interpreting services, especially as these technologies begin to approach the quality of professional interpreters. They require clear, informed guidance on what AI interpreting can deliver today, what may be possible tomorrow, and how best to integrate these tools to serve broader social, business, and institutional goals.

For human interpreters, the challenge is not merely to defend the legacy of the profession but to articulate where, when, and why human expertise remains essential. This includes shaping the standards, expectations, and ethical frameworks that will determine the profession’s future value. However, doing this without fully grasping the AI revolution happening under our eyes is impossible. Ignoring these shifts only hastens the erosion of professional agency. Engaging with them, critically, strategically, and ethically, offers a more sustainable and empowering path forward.

In both arenas, a renewed and informed engagement with AI interpreting is urgently needed. While there is always a risk that the conversation becomes dominated by hype or reactionary resistance, meaningful progress depends on evidence-based debate, transparent evaluation, and interdisciplinary collaboration.

There is much to be done. Research must keep pace with deployment. Pilot programs should be accompanied by robust evaluation frameworks. Regulatory bodies must consider how existing standards apply, or fail to apply, to AI-mediated communication. Most importantly, professional interpreters, especially those working in high-stakes environments such as healthcare, asylum courts, and diplomacy, must be involved in shaping the future.

Ultimately, the goal is not to resist technological change, but to ensure it serves human needs. AI interpreting can extend access, bridge gaps, and assist professionals. But only with thoughtful integration can it enhance, not erode, the values and principles at the core of meaningful communication.

Expand Your Localization Expertise. Subscribe to Our Newsletter!