Alexa, Siri and the new AI chatbots — what's the difference?
You've had a voice assistant for years. Now there are 'AI chatbots'. Aren't they the same thing? Nearly — and the difference explains why one can set a timer and the other can write a speech.

If you’ve had an Alexa on the kitchen counter since 2018, or you’ve been asking Siri for the weather for a decade, you might reasonably wonder what all the fuss is about. You’ve had a talking computer. What’s new?
The short answer: the old ones follow commands. The new ones hold conversations. That sounds like a small difference. It isn’t.
The old kind: assistants that follow orders
Alexa, Siri, and the original Google Assistant were built around a list of things they can do: set a timer, play music, tell you the weather, turn on a lamp, call your daughter. When you speak, they try to match your words to something on that list.
If your request is on the list, it works beautifully. “Set a timer for ten minutes” — done. If it isn’t, you get “Sorry, I don’t understand that,” or a web search read aloud.
They’re reliable, fast, and limited. They don’t really understand you; they recognise you.
The new kind: chatbots that generate
ChatGPT, Gemini, Copilot and their cousins work the other way round. There’s no list. You say (or type) anything, and they compose a response, word by word, based on patterns learned from an enormous amount of human writing. (How that works.)
So you can ask something nobody has ever asked before — “write a toast for my sister’s retirement that mentions her terrible parking” — and get something usable. You can ask a follow-up. You can say “less formal” and it adjusts. It’s a conversation.
The trade-off: they don’t have a list of reliable actions. Ask ChatGPT to turn on the lamp and it’ll explain how lamps work. Ask it what time it is and it may not know. And because it composes rather than looks up, it can be confidently wrong in a way Alexa’s timer never is.
Which is “better”?
Neither. They’re different tools:
| You want to… | Use… |
|---|---|
| Set a timer, play a song, turn on a light | The assistant (Alexa, Siri) |
| Know the weather, the time, a quick fact | Either — the assistant is faster |
| Write, draft, explain, brainstorm | The chatbot |
| Understand a letter or a document | The chatbot |
| Have a back-and-forth about something | The chatbot |
| Be sure the answer is exactly right | Neither — check a real source |
The two are merging
This is why it’s confusing right now. The companies are putting the new engines into the old assistants. Amazon has upgraded Alexa with a chatbot brain. Apple has been doing the same with Siri. Google has replaced the Assistant with Gemini on many phones.
So the Alexa on your counter may one day — or already — be able to hold a conversation, not just take orders. The tell is simple: ask it something off-list, like “what should I cook with leftover chicken and half a cabbage?” If it answers in full sentences with actual suggestions, it’s got the new brain. If it offers a web search, it hasn’t yet.
Three things worth knowing
Voice assistants listen for their name. Alexa and Siri are always listening for the wake word — that’s how they hear you. They’re not meant to record everything, but they do occasionally mishear and start recording. If that bothers you, there’s a mute button on the device, and you can review and delete recordings in the app.
Chatbots keep a written record. Every conversation is saved in a list you can see, which is convenient — and worth remembering when you’re about to type something private. The usual list applies.
You don’t have to choose. Keep the assistant for the kitchen timer; use a chatbot on the laptop for the letter. That’s what most people end up doing, and it’s fine.
The one-sentence version
An assistant recognises your request and does it if it’s on the list; a chatbot understands your request and writes a response — which is why Alexa is better at timers and ChatGPT is better at speeches, and why they’re slowly becoming the same thing.


