Using Bing's AI chat to play a word association game in French.

AI Chat Prompts for Language Learning Practice

As AI becomes more and more a fixture of daily life, it’s not surprising to see it sneaking gradually into the language learning setting. The list of AI-infused apps is growing daily: premium chatbots, word games and other practice tools that are taking app stores by storm.

But before you shell out the cash, be reassured that you don’t need to spend a penny to bring the magic of AI to your own routine. You can achieve exactly the same, routine-transforming effects with a few handy prompts and free-to-access AI platforms.

AI for Free

There is an impressive and ever-growing array of AI chatbots to experiment with, some free, some premium. The difference, largely, is in the amount of environment setup that has already been done in premium apps. This is often just a case of configuring the AI to play a certain role, or act in a certain way. But here’s the secret:

You can do this easily yourself, with no development skills required.

It’s simply a case of good prompting. Tell it, in simple, natural English, the rules of the activities you want to run. Define the way you want it to respond to you. Give it a role to play. And state your limits and boundaries. Get a handle on prompting effectively, and popular, free platforms will more than suffice.

The biggie, of course, is ChatGPT, which has a solid free tier for general use. Google have also joined the game with the commendable Bard, which is definitely worth a look. That said, since Microsoft released their new AI-powered Bing chat mode, I’ve been using that more and more. It has an excellent Creative Mode preset, which gives it more unpredictable, humanlike responses. Just bear in mind that Bing currently limits the conversation to a 30-interaction maximum.

The following examples use French and Swedish to show how I’ve been using it to support my own language practice. Just swap in your own target language as required!

Word Games

The simplest kind of game to set up is good old basic word play. The following prompt sets up a turn-based alphabet game, which challenges your vocab recall:

Let’s play a word game in French. We have to go through the alphabet in turns, stating a verb that begins with each letter, plus a short sentence using that verb in context. Let’s play!


One of the best things about using AI for these kinds of language games is the capacity for on-the-fly correction and feedback; it can sometimes appear almost human. Below, I started cheating by inventing words, but Bing was far too clever to be caught out!

Playing a language learning word game - and trying to cheat - with Microsoft Bing's AI-powered chat mode.

Playing a language learning word game – and trying to cheat – with Microsoft Bing’s AI-powered chat mode.

Another fun vocab item practice mode is word association. The following prompt sets up a game where the meaning of each turn’s word must be related to the last. If the bot considers the link too tenuous, you lose the game:

Let’s play a word association game in French. You kick us off with a random noun. We then take it in turns to give a word which is somehow related in meaning to the last. If the link is too tenuous, the player loses. Shall we play?


It’s a great way to recycle vocabulary. You might need to play with the prompt to make your AI teacher a little less strict, though. Mine ended up with a bit of a mean streak. Very harsh!

Bing AI being VERY harsh on me in a word association game.

Bing AI being VERY harsh on me in a word association game.

Story Games (with Tutor Mode!)

When you’re ready to take it beyond words, AI is ready for you. One of the most amusing ways to practise with full sentences is storytelling. Try this prompt for a narrative whirl:

Let’s play a turn-based storytelling game to help me practise my French. We build a story by taking it in turns to add a sentence each time. Please keep the language level to about A2, and tell me about any mistakes I make as we go along. The story should be set in the present day. You start us off!


Note the specification of a language level, as well as the instruction to correct your mistakes as you go. It makes the AI response so rich and helpful that it really is a gift to learn from. To tailor it further, try adding instructions about which tenses to use (narrative present or past?), and even vocabulary topics to crowbar in.

Role-play

Ai chat can prepare you for real-world chat, too. Setting up a foreign language role-play is as simple as describing the situation in as much detail as you like:

Let’s do some role-play to help me practise my French! You play a friendly waiter in a Paris café, and I am a customer. I enter the café and you come over to take my order. You realise I’m learning French and so give me very simple descriptions of all the dishes. But you keep mishearing me, so I have to repeatedly rephrase what I ask for.


This can be as straight-laced or as silly as you like. Sometimes, it’s a case of the crazier, the better. There’s nothing like a bit of silliness to increase engagement and recall.

You can even target the chat more by priming the AI with the actual vocabulary items you want to lever in. It’s a great way to recycle words over and over again:

I want to practise talking in Swedish about family. Imagine you’re a friend of mine and we’re having a chat about our families. Keep the language level to about A2 on the CEFR language scale, and using the following words as much as possible: mamma, pappa, bror, syster, vänner, snäll, vänlig, lita på, besöka


Whether you’re new to AI, or just beginning to experiment with it yourself, I hope these sample prompts give you some useful, fun practice ideas. Do you have any good ones to add to the list? Let us know in the comments!

ChatGPT screenshot

ChatGPT for Language Learners

The buzz around AI imaging seems only five minutes ago, yet there’s another brand new tool creating ripples. And this time, it speaks.

ChatGTP is an AI model that processes natural language, making sense of instructions and carrying them out. You could think of it as a kind of ask me anything bot, and it went truly viral at the end of last year thanks to its uncannily human-like language abilities

Of course, it didn’t take long for the language community to see the potential. The algorithm has already captured the imaginations of teachers, who are using it to great time-saving effect in generating quick and simple lesson plans. No surprise, then, that the polyglot community has followed suit in exploring the new tech’s potential for supporting language learning.

As with all tech, the best way to assess it for yourself is to get your hands dirty. In that spirit, I headed to OpenAI.com’s ChatGPT portal to see for myself what it could do. Note that this might be easier said than done right now; lately, you’re more likely to see the message ChatGPT is at capacity right now as the fellow curious inundate the platform with requests.

ChatGPT for (Language) Beginners

I started off at the place that seemed most fitting: at the beginning. What about some learning tips for a newcomer to a specific language, for a specific purpose? ChatGPT turned out solid phrase lists, and – impressively – not always the most obvious cut-and-paste choices. Accompanying advice was on the whole quite generic, but very sensible and practical:

ChatGPT screenshot

What I love is the variability; ask the same question twice, and you’re unlikely to get the same answer. There’s always some overlap, but it’s interesting to see how suggestions vary from answer to answer:

ChatGPT screenshot

Occasionally, you get a bit of extra advice for free, too:

A screenshot of a conversation where the user asks the AI engine ChatGPT for French tips for a trip to France.

ChatGPT seems really good at making what we might call potted lessons like these, which explains its popularity as a quick lesson plan generator.

Off the Beaten Path

Where it struggles, I found, was when you stray from the mainstream path – presumably, fields where the algorithm finds much scarcer material to work with. For example, Explain how tense works with Modern Hebrew verbs produced a very convincing piece of text that sounded like it came straight from a Routledge Comprehensive Grammar. Unfortunately, the Hebrew itself was an absolute hash, omitting any mention of vowel patterns, and focusing on suffixes, as if Hebrew were a typical Romance language or similar.

The problem, I’m guessing, is a paucity of sources. I’m not sure where it cobbled the points together from, but they seemed like a very bad, rookie guess at how to express tense in Hebrew, based on a very limited set of observations. Perhaps I’m being harsh; experimenting with different question phrasing might have improved things, and I’m impressed enough that it dealt so well with Greek.

It’s early days, though. Development is entering a new stage, backed by some big money, and refinements will come thick and fast. Crucially, the spark is already lit; ChatGPT has captured imaginations, and it already looks like a truly helpful and practical tool is emerging. 

Have you taken your first steps with ChatGPT as a language learner? Let us know how you got on in the comments!

An array of neon signs of nonsense words on a wall. Image generated by the Stable Diffusion AI algorithm.

Polyglot in the Machine: AI for Language Learners

AI is the order of the day lately. Have you seen how many fantasy photos have been filling up Instagram lately? Thanks to the now wide availability of open source AI algorithms, some powerful computing power is in the hands of users courtesy of apps like Dawn AI and Lensa. Type in a few words, and the computer does the painting.

It’s new tech, opening new possibilities alongside new ethical challenges that users are gradually becoming sensitive to. But the benefit to individual language learners here is apparent very little imagination stretch. First and foremost, these algorithms parse human language. So why not, for instance, type in some target language – say, ein Hund mit grünen Augen (a dog with green eyes) – and see if the picture matches what you meant to say? It should act as a kind of machine validation that the language you produce makes sense.

It already works to a point with some languages. Models like Dall-E (seen at work below in the web-based Craiyon.com) cope reasonably well with non-complex, non-English prompts.

A screenshot from Craiyon.com, a web-based AI image generator built on DALL-E Mini.

It can be hit and miss, but Craiyon understood my German for the most part!

So it works – up to a point. The current stumbling block is linguistic and cultural bias. For a start, models like Stable Diffusion were initially developed and trained with English input. And as one web experimenter shows, non-English results can leave a lot to be desired, with a definite advantage for Western European languages. This isn’t surprising, given that the technique samples from pre-existing web content; the predominance of certain languages means there is a lot more of that to learn from.

Ai Work In Progress

It’s clear these techniques are nascent and emerging, as most casual users will admit. Even if English is your target learning language, for example, images can frequently be so off the mark that you may question whether it understood a single word of your prompt.

Things are improving, though, especially with regular updates to the Stable Diffusion model. There are even a couple of language augmentation projects floating around in beta, including. one that adds ‘Japanglish’ capabilities to the current algorithm, overcoming one particular cultural blindspot.

And, if you have the skills, you can add to many ongoing open source projects to extend and finesse the capabilities of AI algorithms. I’m sad to report that that isn’t in my skillset, but it’ll be interesting to follow how this develops over the coming months!

Amazon Echo Dot - Alexa for Language Learning

Alexa: Your Personal, Digital Native Speaker

It’s a language learning ‘secret’ that isn’t so secret any more: changing the language setting on your smart devices is a brilliant way to create a personalised immersion environment without going abroad. And the recent explosion of artificially intelligent digital assistant devices is taking this one step further. Voice-activated gadgets, like Amazon’s Alexa, place a (robotic) native speaker right in the centre of your home.

Swayed by the temptingly low price on the entry-level Amazon Dot, I’ve been getting to know Alexa for the past few months. First off, it’s a cliché, but this is definitely the kind of gadget you ‘never knew you needed’. After eyeing the unit with some cynicism for the first few weeks, soon I was constantly asking it to play music, convert currencies and measurements, tell me the weather forecast or simply the time. It’s both easy and fun, and gives you that sense of the future is now!

You digital language assistant

But it’s not just about voice-activating mundane, daily tasks. Ever alert to new learning opportunities, changing Alexa’s language settings was top of the list of experiments to try. And it works a treat, especially for pronunciation; suddenly, I was having to focus intently on expressing my commands in a nice, clear German accent so that Alexa could understand. (Incidentally, I’ve also found switching the language of Apple’s assistant Siri has these great pronunciation drill benefits!)

Interacting is as simple as asking a question like “”Alexa, was sind die Nachrichten?” (Alexa, what’s the news?) or “Alexa, wie ist das Wetter heute?” (Alexa, what’s the weather like today?). For more capabilities – including lots of silly (but briefly entertaining) games – there are hundreds of extra installable skills on Amazon. A useful hit list of the most useful can be found here.

The only snag with Alexa is that it is currently only available in English or German. Great news for Germanists, who won’t feel underrepresented in the language learning world for a change; but a pretty large black hole for everyone else.

Skilling up Alexa as language tutor

However, all is not lost. Users can still download Alexa Skills from Amazon, which augment the device’s capabilities. Already there are a good number of language learning skills, although they vary greatly in quality. It’s clearly early days for the device in terms of educational skills, but the start is promising.

A simple search on Learn Spanish or similar will yield plenty of results for you to try out. Here are a couple of links for the more mainstream languages:

Feedback ranges from decent right down to downright terrible on some of the skills available. However, the facility to give feedback on Amazon is a route for users to shape and improve Alexa as a language learning tool. Try new skills out, and write an honest review for each one – your thoughts will help developers to tweak and adapt Alexa skills for an incrementally better experience.

Watch this space

In summary, Alexa is an excellent investment for Germanists, but hit and miss for students of other languages – at least for the time being. There is a sizeable clamour around Spanish support on Amazon’s developer space, with pressure for other languages too. It would only seem a matter of time before she becomes more than just bilingual.