The importance of the human element in language learning

The research we’ve done at Berlitz shows that people prefer to learn from other people. Even though AI can be a great language learning tool, it still can’t fully adapt to a learner’s preferences, emotions, and pace. And, most of all, it lacks the human-to-human interaction, which (research shows) is one of the primary motivation factors for language learners around the world.

Social learning: The power of connection

One thing we all need to remember about artificial intelligence is that it's not real intelligence. Most AI tools used in language learning are built on massive statistical models that predict the next word, which is essentially a very advanced form of text prediction. What they can’t do is think critically, empathize, or adapt their approach based on a learner’s emotions, body language, or confusion.

This is huge because, beyond simply delivering information, teaching is about reading people. Human teachers can sense a student’s discomfort, celebrate small wins, and change course in real time. They add something AI can’t: storytelling, cultural exchange, and personal anecdotes that make lessons memorable and relatable.

Learning with others amplifies this effect. Being in a classroom with peers, sharing jokes, and building a bond with a teacher dramatically increases motivation and provides students with a tangible sense of progress. Surveys consistently show that while students enjoy using AI chatbots for targeted tasks, they find live, human-led interaction far more motivating and memorable. One study even found that 27.2% of online students’ engagement could be explained by instructors’ emotional intelligence, even in remote settings.

And cultural contexts matter too. In collectivist cultures (like Latin America and Italy), learning is naturally social. In more individualist cultures (like Germany and Poland), motivation often relies on structured accountability from a teacher. Either way, the emotional connection, whether it’s from peers or teachers, drives persistence, which is something AI alone still struggles to replicate.

shutterstock_2324447645.webp

Learning Effectiveness at Stake: Humans vs. AI

AI is now part of almost every stage of language learning. Teachers use it to create grammar and vocabulary exercises (41%) and generate reading or listening materials (35%), while students rely on it for pronunciation practice, instant feedback, and discovering new vocabulary. Yet despite these benefits, about 85% of learners still rate human interaction as “very” or “somewhat” important in their language learning journey.

In many aspects of language teaching, such as building emotional rapport, giving spontaneous corrections, and adapting to cultural nuances, human instructors consistently outperform AI tutors. AI tools might correct grammar, suggest pronunciation drills, or flash new words, but it cannot sense frustration or adjust explanation style to match each learner’s history. And it often gets confused, even with simple tasks, offering the wrong explanations or examples that don’t really make sense.

That is why at Berlitz, we believe success comes from combining human-led teaching with AI-powered personalization. When done well, this approach yields better outcomes, deeper satisfaction, and a more lasting impact than either method alone. And both teachers and students around the world seem to agree with us.

Motivation that powers success

Gamified platforms attract millions of users, but motivation often fades when learners study alone. Research backs this up: an EFL study in Saudi Arabia found that students who took part in group activities and received peer feedback showed significantly higher motivation and persistence than those relying only on apps.

That’s because motivation in language learning is fueled by things AI can’t truly replace, such as peer accountability, interactive activities like debates and role-plays, recognition from teachers and classmates, and even the subtle pressure of not wanting to disappoint someone who cares. AI can mimic praise, but it can’t deliver the empathy in a teacher’s tone, expression, or body language that makes feedback feel meaningful.

When motivation is nurtured, progress accelerates. And progress feeds motivation. Small wins such as speaking more confidently, participating in class debates, or getting positive feedback from a tutor create a loop that keeps learners engaged.

AI limitations & Challenges in Language Learning

Despite its promise, AI in language learning has clear limitations, especially when handling complex tasks or non-English languages. It makes mistakes, struggles with following complicated prompts, and often takes the “easy way out” by inventing things that aren’t true. A recent study shows that LLMs (like GPT-3.5, PaLM, LLaMA) increasingly produce confident but false responses ("hallucinations"), especially when dealing with inference tasks that aren’t well supported by their training data.

Because most large language models are primarily trained in English, AI hits a ceiling when it comes to accurately translating English into other languages, explaining advanced grammatical concepts, and even creating exercises tailored to a specific proficiency level.

In less widely spoken languages, AI frequently produces content that is inaccurate or inconsistent. It also often translates things without cultural context and nuance, which leads to sentences that sound unnatural to native speakers. A study found that a model trained natively in Swahili made about 4 times fewer errors than when using English-trained systems on the same Swahili tasks. This supports the idea that non-English languages suffer when models are mostly trained on English.

And aside from linguistic issues, there’s also the education component. Just because a system has access to information doesn’t mean it can teach effectively. In fact, learners face a paradox: they must know enough to judge whether the AI is teaching correctly. This highlights the biggest gap of all: while AI can deliver information, it can’t observe, respond, or care in the way a human teacher can.

A teacher, tutor, or instructor can read a student’s body language, adjust pace, vary activities, and give personalized feedback in ways AI simply cannot. That’s why at Berlitz, we integrate AI technology without relying on it entirely. It’s a support tool, not a substitute for skilled human guidance.

AI can teach, but Berlitz ensures you learn

Let’s be real: educators and institutions that ignore AI risk being left behind. Used wisely, AI offers real advantages, such as faster progress and increased independence for learners, as well as lighter workloads for instructors.

At Berlitz, we believe the best way to leverage AI is to make it work with us, not instead of us. Our approach blends the speed and scale of AI with the empathy, expertise, and adaptability of trained instructors.

Instead of generic tech, our AI tools are built on decades of language-teaching experience, carefully trained by humans using the Berlitz Method. Every AI feature we release undergoes extensive prompt engineering, testing, and refinement to ensure it adds real value for learners. And when AI can’t be trusted to deliver an acceptable output, we put humans in the loop.

We offer the best of both worlds, and the results we notice among our students confirm that. We clearly see that incorporating AI into the learning experience has led to an increase in learning sessions and student engagement. Most of all, integrating AI features such as the MyBerlitz Speaking Tutor has increased the average speaking time of our learners up to 6 hours per month. Also, 93% of the students we asked agreed that the MyBerlitz Speaking Tutor was very helpful to their learning experience.

How we combine AI and human instruction in our programs

We’ve built AI directly into our online platforms, Live Online classes, Berlitz Flex, and Berlitz On Demand, both of which are accessible on mobile devices. Both programs guide learners along structured, self-paced video lessons and activities that they unlock gradually as they progress. And along the way, learners also gain access to AI-powered tools, such as interactive exercises, quizzes, and games, all directly tied to the topics they’re studying.

Our most powerful AI-driven feature is the MyBerlitz Speaking Tutor, an AI features trained on the proven Berlitz Method. This features enables learners to practice speaking and pronunciation with instant, real-time feedback. Currently available for four of our core languages (English, French, German, and Spanish), it simulates real-world scenarios as closely as possible, allowing learners to choose voices and accents and even rehearse job-specific topics such as presentations, meetings, or interviews.

Aside from that, learners also have access to the AI Learning Assistant, a 24/7 chatbot that supports language development on demand. It can answer questions, provide explanations, translate, proofread emails or presentations, and offer feedback exactly when needed.

None of our AI features are generic add-ons bolted onto a course. They’re built around learning paths designed by Berlitz instructors, learning designers, and grounded in the Berlitz Method. That means everything our AI does reinforces immersive, target-language learning focused on speaking and real-life vocabulary. And because we know that unstructured chatbot conversations can be mentally draining, we’re deliberate about keeping AI at the learner’s level and weaving it into a balanced mix of activities.

What’s next: Our vision for AI-powered learning

For now, we’re only scratching the surface of what AI can do for language learning. Our next chapter is about building an AI ecosystem that supports both students and instructors by personalizing learning paths and aligning AI more closely with real-world scenarios. In the future we’re envisioning, our AI will be able to generate personalized lessons and materials on the fly. Instead of cookie-cutter curricula, each student could follow a course built just for them. A built-in co-pilot will guide learners like a GPS for their language journey, suggesting what to review next. Ideally, our AI will amplify, not replace, human connection with personalized reminders, motivational check-ins, and feedback loops that feel genuinely supportive, even though they’re powered by technology.

Naturally, our AI Speaking Tutor will evolve as well. First, it will expand to more languages. Second, it will allow learners to customize how their tutor interacts with them (for example, choosing between real-time feedback or feedback at the end of the conversation). Third, it will adapt to real-life contexts and cultural nuances (Mexican Spanish vs. Castilian Spanish).

We’re also exploring smarter onboarding and retention flows to help students stay motivated, plus automated quality checks that transcribe lessons, measure participation, and ensure the Berlitz Method is applied correctly. These tools will give instructors and managers real-time insight into what’s happening in every lesson so they can act faster to improve outcomes.

And on the instructor side, AI assistants will help prepare materials, suggest activities, and flag student needs in real time, freeing instructors to do what they do best: teach. AI will also help reduce the time and cost of training new instructors, modeling lessons, giving feedback, and speeding up the certification process.

"We know what truly drives progress"

After 147 years teaching languages, we know what truly drives progress. That’s why each of our AI tools is trained for one clear task and always grounded in the Berlitz Method, our proven approach to immersive, goal-based learning. Our MyBerlitz Speaking Tutor, for example, focuses entirely on helping learners reach their speaking goals after every lesson, just like every Berlitz class.”

Jon Hritz, Director of Global Product

Empowering learners through AI and people

At Berlitz, we’re designing learning experiences where AI-powered personalization and human instruction work hand in hand. Separately, each has strengths and limits, but together they create a richer, more effective way to learn.

Even as AI technology advances rapidly and specialized models trained on language-learning data emerge, one thing remains clear: AI still doesn’t replace the empathy, adaptability, and human connection of a skilled instructor. That’s why our strategy is to keep evolving our Generative AI features to support and complement our instructor-led and self-paced programs, not replace them.