Brief answer: Siri is AI because it uses speech recognition, natural language processing, machine learning and automation to understand requests and complete tasks. It is best understood as voice-based assistant AI, not a generative chatbot, especially when you expect long conversations, creative writing or complex reasoning.
Key takeaways:
AI category: Treat Siri as narrow assistant AI, not human-like intelligence.
User expectations: Compare Siri fairly when judging simple commands against open-ended generation.
Task fit: Use Siri for timers, calls, reminders, music and device control.
Limitations: Expect weaker results with ambiguity, follow-ups and multi-step reasoning.
Privacy impact: A stronger Siri needs clearer controls over personal context and permissions.

Articles you may like to read after this one:
🔗 What is Gemini AI?
Learn how Gemini AI supports search, writing, and productivity.
🔗 What is Grok AI?
Discover how Grok AI delivers conversational answers and insights.
🔗 What is Manus AI?
Explore Manus AI and its role in autonomous task automation.
🔗 What is OpenClaw AI?
Understand OpenClaw AI and how it supports open-source agents.
🔗 What is AI-powered search?
See how AI-powered search improves discovery, answers, and research.
1. What Does “AI” Mean? 🧠
Before judging Siri, it helps to untangle the word AI, because people throw it around like confetti at a tech conference.
Artificial intelligence is not one single thing. It is a broad category of technologies designed to make computers perform tasks that normally require human-like intelligence. That can include:
-
Understanding speech
-
Recognizing images
-
Predicting what a user wants
-
Translating languages
-
Answering questions
-
Sorting emails
-
Recommending music
-
Driving cars, sort of
-
Generating text, images, or code
-
Making decisions based on patterns
So AI does not always mean a robot with opinions and a glowing blue brain. It can be something much quieter, like your phone predicting the next word you’ll type or your camera improving a photo in the background. That kind of AI is everywhere now, even when it doesn’t announce itself with dramatic sci-fi lighting.
Siri fits inside this broad AI category because it uses intelligent systems to process your voice, interpret your request, and trigger a response or action.
When you say, “Hey Siri, set a timer for ten minutes,” Siri has to do several things:
-
Detect the wake phrase
-
Capture your speech
-
Convert your voice into text
-
Figure out your intent
-
Identify the relevant app or system function
-
Set the timer
-
Confirm the action
That might feel simple because it happens quickly. But under the hood, it is a neat little sandwich of AI systems, automation, and software engineering. Slightly dry sandwich maybe, but still.
2. Is Siri AI or Just Voice Recognition? 🎙️
This is one of the biggest points of confusion. Siri is not only voice recognition.
Voice recognition is part of Siri, but Siri also needs to understand meaning. That’s where natural language processing comes in.
For example, these commands all point toward a similar action:
-
“Set an alarm for seven.”
-
“Wake me up at seven.”
-
“I need to get up at seven.”
-
“Make sure I’m awake by seven.”
A basic voice recognition system might only transcribe the words. Siri has to go one step further and infer that you want an alarm. That inference is an AI task.
Now, is Siri perfect at this? Absolutely not. Sometimes it hears “call Mum” and decides you meant “open mushrooms” or something equally possessed. You know how it is. Voice assistants can feel magical one second and like a toaster wearing glasses the next. 🍞👓
Still, the fact that Siri can interpret varied language, map it to actions, and improve based on patterns is a clear sign that Siri uses AI.
So, Is Siri AI? Yes - but it is more accurate to say Siri is a voice-based AI assistant, not just a speech-to-text tool.
3. How Siri Uses Artificial Intelligence in Daily Life 📲
Siri’s AI works best when the task is clear and connected to your device or Apple ecosystem. It is not trying to be a philosopher in your pocket, or at least not very convincingly. It is mostly designed to help you get things done.
Here are the main AI-powered functions Siri relies on:
Speech Recognition
Siri listens to your spoken words and converts them into text. This requires models trained to recognize different accents, speaking speeds, background noise, and phrasing.
Natural Language Processing
Once Siri has the words, it tries to understand what they mean. This is the difference between hearing a sentence and understanding the task behind it.
Intent Detection
Siri identifies what you want. Are you asking a question? Sending a message? Controlling a smart light? Starting navigation? Making a reminder? Intent detection is a huge part of assistant AI.
Context Awareness
Siri can use some context from your device, like contacts, location, apps, calendar entries, messages, music preferences, and connected accessories. This helps it respond more helpfully.
Personalization
Siri can become more helpful over time by adapting to repeated habits and preferences. Not in a “it knows your soul” way, hopefully - more like it notices patterns in how you use your phone.
Automation
Siri connects your request to actions. It can launch apps, control settings, start shortcuts, send texts, or manage smart home devices.
This is why Siri is more than a novelty. It is part AI, part interface, part automation layer, and part digital butler that sometimes drops the tray.
4. Why People Still Ask “Is Siri AI?” 🤔
People keep asking “Is Siri AI?” because the public idea of AI has changed dramatically. For a long time, Siri, Alexa, and Google Assistant were some of the most visible examples of consumer AI. They talked, responded, and handled requests. That felt futuristic.
Then conversational AI tools changed what people expected from artificial intelligence.
Now people expect AI to:
-
Write essays
-
Summarize documents
-
Explain complex topics
-
Generate ideas
-
Remember context across a conversation
-
Help with coding
-
Analyze images
-
Draft emails
-
Rewrite content in a certain tone
-
Reason through multi-step problems
Compared with that, traditional Siri can feel narrow. It may be AI, but it does not always feel like the polished new version of AI people are thinking about.
That’s the tricky part. Siri is AI by definition, but it may not match the modern mental image of AI.
It’s like asking whether a bicycle is transportation after seeing a high-speed train. Yes, obviously it is transportation. But the experience is not the same, and pretending otherwise would be silly.
5. Siri AI vs Generative AI: What’s the Difference? ✍️
This is where things get important.
Siri is mainly an AI assistant built for command handling and device control. Generative AI is designed to create new content, such as text, images, audio, code, summaries, and more.
Siri traditionally works best with short commands:
-
“Text Alex.”
-
“Play my playlist.”
-
“Turn off the kitchen lights.”
-
“Remind me to buy coffee.”
-
“What’s the weather?”
-
“Call Dad.”
Generative AI works better with open-ended requests:
-
“Write a meal plan for a busy parent.”
-
“Explain quantum computing like I’m tired.”
-
“Draft a polite complaint email.”
-
“Compare three business ideas.”
-
“Brainstorm names for a podcast.”
-
“Summarize this long document.”
That difference matters because users now expect assistants to be flexible, conversational, and context-aware. Siri’s older assistant model can feel rigid by comparison.
So yes, Is Siri AI? Definitely. But Siri and generative AI are not identical categories.
Siri is more like a command center. Generative AI is more like a thinking-and-writing partner. Sometimes they overlap, but they were not originally built around the same main purpose.
6. Comparison Table: Siri AI vs Other Types of AI Assistants 📊
| AI type / assistant style | Best for | How it feels to use | Strengths | Weak spots |
|---|---|---|---|---|
| Siri AI | Voice commands, phone tasks, smart home controls | Fast, simple, sometimes stubborn in peculiar ways | Hands-free control, device integration, reminders, calls | Not always great at long conversations |
| Chatbot-style AI | Writing, explaining, brainstorming | More conversational, like texting a brainy intern | Flexible answers, content generation, deeper responses | May need careful prompting |
| Smart home assistant | Lights, speakers, thermostats, routines | Practical, household-ish | Great for connected devices | Limited outside home commands |
| Search assistant | Quick facts, directions, local info | Search box with a voice | Good for finding info fast | Often less creative or personal |
| Automation tools | Repeated tasks and workflows | Quiet magic in the background ✨ | Saves time, links apps together | Setup can feel fiddly |
| Predictive AI | Recommendations, autocorrect, suggestions | Invisible until it annoys you | Personalization, speed | Can guess wrong in confident fashion |
This table shows why the phrase Siri AI can feel both accurate and incomplete. Siri is absolutely an AI assistant, but it occupies a specific lane. It is not trying to be every kind of AI at once, though users increasingly expect exactly that.
7. What Siri AI Is Good At ✅
Siri’s strengths are easy to overlook because many of them are small. But small tasks matter. They’re most of life.
Siri is helpful when your hands are busy, your screen is locked, or you need to do something quickly without opening an app.
Siri is good at:
-
Setting timers while cooking 🍳
-
Creating reminders
-
Sending quick messages
-
Starting phone calls
-
Controlling music
-
Checking weather
-
Opening apps
-
Starting navigation
-
Managing calendar items
-
Controlling smart home devices
-
Running shortcuts and routines
-
Reading notifications aloud
-
Doing basic calculations
These are not glamorous tasks. Nobody writes a movie where the hero says, “Set a pasta timer,” and the audience gasps. But that convenience adds up.
Siri’s value is practical ease. It reduces friction. Instead of tapping through menus, you speak naturally and let the system do the dull bit.
That’s a valid AI use case. Not all artificial intelligence has to write poetry about moonlight and capitalism. Sometimes AI just needs to turn on the hallway light without making you stand up.
8. Where Siri AI Falls Short 😬
Now for the slightly awkward part. Siri has limitations, and users notice them.
The most common complaints are not usually about whether Siri is AI. They’re about whether Siri is good enough AI.
Siri can struggle with:
-
Multi-step questions
-
Follow-up context
-
Ambiguous wording
-
Complex reasoning
-
Long explanations
-
Open-ended creative requests
-
Understanding unusual names
-
Handling noisy environments
-
Doing tasks across non-native apps
-
Remembering conversational context
Sometimes Siri gives a web-style answer when you wanted an action. Sometimes it performs an action when you wanted information. Sometimes it says it cannot do something that feels like it should be obvious. And sometimes it just spins in place like a shopping cart with one cursed wheel. 🛒
That does not mean Siri is not AI. It means Siri’s AI design has historically been optimized for assistant tasks rather than broad reasoning.
This is a key distinction. A calculator app is still software even if it cannot edit photos. Siri is still AI even if it cannot produce a polished business strategy or carry a ten-message conversation without losing the plot.
9. Why Siri Feels Less “Smart” Than Newer AI Tools 🧩
The emotional side of this question matters. People do not judge AI only by its technical definition. They judge it by how it feels.
Modern AI chat systems often feel smart because they generate long, fluent, human-like responses. They can explain things in paragraphs, adjust tone, revise answers, and respond to unruly prompts. That creates the sensation of intelligence.
Siri often gives short responses and depends on predefined categories of tasks. It may understand you, but it usually doesn’t talk through the answer in a rich way. So the interaction can feel less intelligent, even when AI is definitely involved.
There are a few reasons for that:
Siri is voice-first
Voice interactions need to be brief. Nobody wants their phone delivering a six-minute lecture when they asked for tomorrow’s weather. Well, maybe one person does, but they’re probably folding socks by themselves.
Siri is action-first
Siri is designed to do things: call, text, remind, play, navigate, control. Generative AI is designed to produce and reason through content.
Siri is tied to privacy and device behavior
Personal assistants often have to balance practicality with privacy, permissions, safety, and reliability. That can make them more cautious and constrained.
Siri has to avoid risky mistakes
If a chatbot gives a peculiar paragraph, that’s annoying. If a voice assistant sends the wrong message or changes the wrong setting, that can be more disruptive. Assistant AI often needs tighter guardrails.
So Siri may feel less flexible, but some of that limitation comes from what Siri is supposed to do.
10. Is Siri AI in the Same Way a Chatbot Is AI? 💬
Not exactly.
Both Siri and chatbots use AI, but they are built around different interaction models.
Siri is primarily a task assistant. A chatbot is primarily a conversation and content assistant.
Think of Siri as the receptionist who knows how to route calls, manage your schedule, and unlock the meeting room. Think of a chatbot as the colleague who can draft the agenda, summarize the meeting, and explain why everyone is suddenly saying “synergy” again. Neither metaphor is perfect; it wobbles a bit. But it gets the point across.
Siri’s intelligence is practical and command-based. Chatbot intelligence is language-heavy and generative.
That’s why asking “Is Siri AI?” is a bit like asking whether a microwave is kitchen technology. Yes. But it’s not the same as a full commercial kitchen.
11. The Role of Machine Learning in Siri AI 🛠️
Machine learning is one of the core technologies behind Siri AI. It allows systems to detect patterns and improve performance based on data.
For Siri, machine learning can help with:
-
Recognizing speech patterns
-
Improving dictation accuracy
-
Understanding different accents
-
Predicting likely commands
-
Ranking possible interpretations
-
Suggesting relevant actions
-
Personalizing responses
-
Reducing errors over time
For example, when Siri hears a phrase, there may be several possible interpretations. Machine learning helps determine which interpretation is most likely based on language patterns, context, and user behavior.
The system is not “thinking” like a person. It does not sit there pondering your alarm with a tiny cup of coffee. But it does use statistical models and learned patterns to make decisions.
That is AI in the practical, everyday sense.
12. Does Siri Understand You or Just Match Commands? 🧐
This is the philosophical-ish section, but not too philosophical because we all have errands.
Siri does not understand language the way humans do. It does not have personal experience, emotions, common sense in the human sense, or a deep inner model of life. It does not “know” what being late for work feels like.
But Siri can process language well enough to map many spoken requests to helpful actions. That is a form of machine understanding, even if it is not human understanding.
There’s a spectrum:
-
Basic command matching: “Say this phrase, trigger that action.”
-
Natural language understanding: “Interpret different phrases with the same meaning.”
-
Contextual reasoning: “Use previous information and surrounding context.”
-
Generative reasoning: “Create new responses and solve open-ended tasks.”
Siri sits strongest in the first three, depending on the request. Generative AI tools are stronger in the fourth.
So when people say Siri is “not real AI,” they usually mean it does not feel deeply conversational or creative. But technically and functionally, Siri is real AI.
13. What Siri AI Needs to Get Right Going Forward 🚀
For Siri AI to feel more modern, it needs to improve in areas users care about. Not flashy demos. Not awkward party tricks. Genuine everyday assistance.
The big areas are:
Better Conversation Memory
Users want Siri to understand follow-up questions. If you ask about a restaurant, then say “call them,” Siri should know who “them” is. Context is everything.
Deeper App Control
People want Siri to do more inside apps, not just open them. The dream is saying, “Find the photo from my beach trip where everyone is smiling and send it to Jamie,” and having the assistant manage that.
More Natural Language
Siri needs to handle untidy human phrasing better. People don’t speak in perfect command templates. We mumble, backtrack, interrupt ourselves, and say things like, “Can you remind me about that thing later?” Very clear, yes, thanks brain.
Smarter Summaries
A modern Siri AI experience should summarize messages, notifications, emails, and calendar conflicts in helpful ways. Not too much. Just enough.
Better Reliability
This is unglamorous but huge. Users forgive limited features if the assistant is reliable. They get frustrated when basic tasks fail.
Strong Privacy Controls
A personal assistant touches sensitive parts of life: contacts, messages, location, voice, calendar, home devices. People need clear boundaries and control.
In other words, Siri does not just need to become “more AI.” It needs to become more helpful in a way that fits naturally into everyday phone use.
14. So, Is Siri AI or Not? The Clean Answer 🎯
Is Siri AI? Yes. Siri is AI because it uses artificial intelligence technologies to listen, interpret, respond, and act on user requests.
But Siri is not the same as a modern generative AI chatbot. It is better understood as a voice-based digital assistant powered by AI, designed mainly for commands, automation, device control, and quick information.
A good way to put it:
Siri is AI, but it is assistant AI - not necessarily full conversational generative AI.
That distinction clears up most of the confusion.
Siri can be smart in small, practical moments. It can also be frustratingly limited when you expect it to reason, create, or hold a rich conversation. Both things can be true. That’s the whole story.
15. Practical Examples: When Siri AI Makes Sense 🧪
Here are everyday examples where Siri AI is genuinely helpful:
Cooking
“Hey Siri, set a timer for twelve minutes.”
No greasy fingers on the screen. Beautiful. Tiny miracle.
Driving
“Hey Siri, text Sam that I’m on my way.”
Hands-free messaging is one of the clearest practical wins.
Smart Home
“Hey Siri, turn off the living room lights.”
This is where voice control feels genuinely futuristic, even if the future still includes laundry.
Productivity
“Hey Siri, remind me to email the invoice when I get home.”
Location-based reminders are a strong assistant use case.
Accessibility
Voice assistants can make devices easier to use for people who have difficulty tapping, typing, reading small text, or navigating screens.
Quick Information
“Hey Siri, what’s the temperature outside?”
Fast, direct, and low effort.
This is where Siri AI does its best work. It’s not always dramatic, but it is often helpful in the little cracks of the day.
16. Common Misconceptions About Siri AI 🧯
Let’s clear up a few myths.
“Siri is not AI because it makes mistakes.”
AI systems make mistakes all the time. Error does not disqualify something from being AI. In fact, probabilistic systems often make mistakes because they are interpreting patterns rather than following one rigid script.
“Siri is just a search tool.”
Siri can search, but it can also control apps, automate tasks, understand spoken language, and interact with device settings.
“Siri must be conscious if it is AI.”
Nope. AI does not mean consciousness. Siri has no feelings, personal desires, or inner life. It is software using models and rules to perform tasks.
“If Siri can’t chat like newer AI, it isn’t AI.”
That’s like saying a camera is not technology because it cannot make phone calls. Different AI systems serve different purposes.
“Siri hears everything and understands everything.”
Siri’s functionality depends on wake detection, settings, permissions, processing limits, and supported commands. It does not have unlimited understanding.
The mistake is treating AI like a single ladder where every system must climb toward human-like intelligence. In reality, AI is more like a crowded toolbox. Siri is one tool in that box. Maybe not the fanciest tool, but still a tool. 🔧
17. Closing Takeaway: Is Siri AI? Yes - But With an Asterisk ⭐
So, Is Siri AI? Yes, absolutely. Siri is an artificial intelligence voice assistant that uses speech recognition, natural language processing, machine learning, intent detection, and automation to help users interact with devices.
But Siri is not always the kind of AI people now imagine when they hear the term. It is not primarily a creative writing partner, research assistant, coding helper, or deep conversational companion. It is more of a practical digital assistant - built to listen, understand commands, and perform actions.
That makes Siri AI, just not the whole AI universe.
The confusion comes from how fast expectations have changed. A voice assistant once felt like the peak of consumer artificial intelligence. Now people compare it with tools that can draft essays, explain strategy, analyze files, and brainstorm brand names before your coffee cools. ☕ That comparison makes Siri feel smaller.
But smaller does not mean fake.
Siri AI is real AI. It is narrow, task-focused, sometimes clunky, sometimes genuinely helpful, and very much part of the broader artificial intelligence landscape. It may not always blow your mind. But when it sets the timer while your hands are covered in flour, it earns its keep.
Real-world example: Using Siri AI to run a morning routine
Scenario
Imagine a freelance designer who begins most mornings with the same small tasks: checking the weather, reviewing the first calendar appointment, turning on the desk lamp, starting a focus playlist, and setting a reminder to send invoices at 4pm.
None of these tasks is difficult. The problem is repetition. Each one takes only 20-40 seconds, but together they create a small drag before the working day has properly begun.
This is where Siri AI makes practical sense. It is not being used as a deep reasoning assistant. It is being used as a voice-controlled automation layer for predictable, everyday actions.
What Siri needs
To make this valuable, the user would need:
A clear Siri phrase, such as “Start my work morning”
Calendar access enabled
Reminders access enabled
Music or podcast app access, if needed
Smart home access, if using lights or plugs
A simple Shortcut that runs the same actions in order
A manual review habit for anything important, such as client messages or payments
Example instruction
Create a Siri Shortcut called “Start my work morning”.
When I say “Hey Siri, start my work morning”, do the following:
Tell me today’s weather.
Tell me my first calendar event.
Turn on my desk lamp.
Start my focus playlist.
Set a reminder for 4pm that says “Send client invoices”.
Do not send messages, emails, or payments without asking me first.
This is a strong Siri-style use case because the tasks are specific, repeatable, and action-based. A weaker instruction would be something vague like “Help me be productive today”, because Siri may not know which actions to take or which apps to use.
How to test it
Run the shortcut for five working mornings and check three things:
Did Siri trigger the correct routine every time?
Did each step complete correctly?
Did any action need manual correction?
A simple test log could look like this:
Day 1: 5/5 actions completed correctly
Day 2: 5/5 actions completed correctly
Day 3: 4/5 actions completed correctly because the playlist app did not open
Day 4: 5/5 actions completed correctly
Day 5: 5/5 actions completed correctly
That gives the user a practical reliability score instead of a vague feeling that Siri is either “smart” or “unhelpful”.
Result
Illustrative result: based on timing five sample mornings, the manual version took about 3 minutes 20 seconds per day. The Siri Shortcut version took about 35 seconds, including listening to the spoken responses.
That saves roughly 2 minutes 45 seconds per morning, or about 13 minutes 45 seconds across a five-day work week.
Accuracy in this example test was 24 completed actions out of 25, giving a 96% completion rate. The one failed action was not a language-understanding failure; it was an app playback issue. That distinction matters because Siri AI can understand the request correctly and still fail if the connected app, permission, or device state is not ready.
What can go wrong
The biggest mistake is asking Siri to handle tasks that need judgement without human review. Turning on a lamp is low risk. Sending a client email, deleting files, or paying a bill is not.
Other common problems include vague shortcut names, missing app permissions, unreliable smart home devices, background noise, and commands that sound too similar to each other.
For important workflows, Siri should confirm before taking irreversible actions. The safest setup is: let Siri prepare, remind, open, or summarise; let the human approve, send, delete, or pay.
Practical takeaway
Siri AI is most valuable when the job is clear, repeatable, and tied to device actions. It does not need to behave like a chatbot to matter. In this kind of routine, its strength is reducing small daily taps, checks, and app-switching into one spoken command.
FAQ
Is Siri AI or just a voice assistant?
Yes, Siri is AI because it uses artificial intelligence technologies to understand speech, interpret requests, and carry out actions. It is more accurate to describe Siri as a voice-based AI assistant than as a simple voice tool. Siri combines speech recognition, natural language processing, intent detection, machine learning, and automation to respond to everyday commands.
Why does Siri feel less smart than newer AI chatbots?
Siri can feel less smart because it was built mainly for quick voice commands and device control, not extended conversations or creative tasks. Newer chatbot-style AI tools often generate detailed answers, summarize text, brainstorm ideas, and sustain richer back-and-forth exchanges. Siri is still AI, but it usually works in a narrower, more task-focused way.
What kind of AI does Siri use?
Siri uses several types of AI working together. These include speech recognition to convert your voice into text, natural language processing to interpret meaning, and intent detection to work out what action you want. Machine learning also helps Siri recognize patterns, improve responses, and make better guesses based on context and repeated use.
Is Siri AI the same as generative AI?
Siri AI is not the same as generative AI in the usual sense. Siri is mostly designed to handle commands, control device features, send messages, set reminders, and answer quick questions. Generative AI is better suited to creating text, summaries, ideas, images, code, and longer explanations. The two can overlap, but they are built around different main purposes.
What is Siri best used for?
Siri is best used for short, practical tasks where voice control saves time. Common examples include setting timers, creating reminders, sending quick texts, starting calls, checking the weather, opening apps, playing music, controlling smart home devices, and running shortcuts. Its strength is convenience, especially when your hands are busy or you do not want to tap through menus.
Why does Siri sometimes misunderstand commands?
Siri can misunderstand commands because spoken language is highly varied and affected by accents, background noise, phrasing, and context. It has to decide both what you said and what you meant, which can lead to mistakes. A common issue is when Siri hears the words correctly but maps them to the wrong action or fails to understand the intent.
Does Siri understand language like a human?
No, Siri does not understand language the way a human does. It does not have emotions, personal experience, or human common sense. Instead, it processes language patterns and maps many spoken requests to likely actions. That is still a form of AI-powered language understanding, but it is not the same as human understanding or conscious thought.
How is Siri different from a chatbot?
Siri is mainly a task assistant, while a chatbot is usually a conversation and content assistant. Siri is strongest when you ask it to do something, such as call someone, set a reminder, or control a device. Chatbots are stronger at explaining ideas, drafting text, brainstorming, summarizing, and handling open-ended prompts that require longer responses.
Is Siri real AI if it makes mistakes?
Yes, Siri is still real AI even when it makes mistakes. AI systems often work by interpreting patterns and probabilities, which means they can misunderstand speech, choose the wrong intent, or fail with ambiguous requests. Making errors does not mean Siri is not AI. It means Siri’s AI is limited and optimized for specific assistant-style tasks.
What would make Siri AI more helpful in the future?
Siri AI would feel more helpful with better conversation memory, stronger follow-up understanding, deeper app control, more natural language handling, smarter summaries, and improved reliability. Users usually want practical improvements rather than flashy tricks. A more helpful Siri would understand context better, complete tasks inside apps more smoothly, and remain dependable for everyday phone use.
References
-
Apple - Siri - apple.com
-
IBM - Artificial Intelligence - ibm.com
-
Apple Machine Learning Research - Hey Siri - machinelearning.apple.com