If you teach languages, you've probably been pitched at least one AI tool that promised to change everything about your classroom. Maybe you tried it. Maybe it landed flat. Maybe it actively got in the way of what you were trying to do.
That experience is real, and it's the reason most teachers we talk to are skeptical of edtech AI. They should be. Most AI tools are aimed at replacing teachers or isolating learners — not at making either one more effective.
There is a useful version of technology in the language classroom, though. The principle behind it is simple: technology that handles the busywork so facilitators can lead real conversations. Used that way, it can give you back hours of your week.
What technology is actually good for in language teaching
The real strength of AI in education isn't replacing what teachers do well. It's doing the things teachers don't have time to do at scale. Specifically:
- Differentiating practice. Generating multiple versions of an activity at different proficiency levels, so each learner gets practice that fits their level.
- Personalizing feedback. Reading a learner's written or spoken output and pointing out specific patterns, both strengths and gaps.
- Building from your context. Taking the topic you're teaching this week and generating prompts, scenarios, or comprehension questions aligned to it.
- Tracking what you can't see at once. Watching where each of your thirty learners is across four skills, so you walk into class knowing who needs what.
What AI is bad at
It's also worth being honest about what AI can't do:
- Read the room. AI doesn't know that a learner's parents are going through a divorce, or that the class is losing focus because the heater broke. You do.
- Run a real conversation. Real conversation requires shared context, cultural reading, and emotional presence. AI gets there occasionally, but a teacher gets there every day.
- Replace judgment. AI can flag that a learner is struggling. It can't tell you whether to push them, slow down, or talk to their counselor. That's still your call.
How to evaluate an AI tool
If a tool is being pitched to you, here are five questions worth asking:
- Does it work with what I already teach? If the tool needs you to abandon your curriculum to use it, that's a red flag.
- Does it save me time, or just give me more data to look at? Information that doesn't change what you do isn't valuable.
- Does it respect my judgment as the teacher? If the tool makes pedagogical decisions over your head, walk away.
- Can I see real classroom data, not demos? Demos are designed for sales. Classroom data tells you the truth.
- What does it do when it's wrong? Every AI is wrong sometimes. Does the tool surface uncertainty, or does it confidently mislead?
Where SoWo lands
We're building SoWo around the principle that technology should make learning more human, not less. AI handles the differentiation. AI surfaces what you can't see at a glance. AI generates the practice each learner needs. But the teacher runs the classroom, makes the calls, and stays in charge of the relationship.
We're co-creating SoWo with OWL-trained teachers because that's the only honest way to build something that works in real classrooms. If you're tired of AI pitches that promise too much, we'd like to be the one that doesn't.