1 of 40

Say

What?

A Journey through Voice and Sound

With Artificial Intelligence

2 of 40

Important Links - Lesson Plan + Video

To save a copy of the presentation:

select from the top menu "File" > "Make a copy" > "Entire presentation".�Choose the location on your computer where the copy will be saved.

3 of 40

4 of 40

What Did We See In The Video?

Discussion

5 of 40

The Application We Saw Used Artificial Intelligence. How Did Artificial Intelligence Help People With Speech Impairment?

Discussion

6 of 40

How does AI help people with speech impairment?

It learns how to “understand” the person with the speech impairment.
It hears the impaired speech, understands it and transcribes it into text.
It created a clear, electronic voice that reads out loud what it understood from the person’s speech.

7 of 40

What Is Artificial Intelligence?

Discussion

8 of 40

What Is Artificial Intelligence (AI)?

A computer’s ability to perform tasks that would normally require human abilities, such as learning, thinking and decision making.

Definition

9 of 40

Speech-to-Text
Text-to-Speech

What Types Of AI Did Project Relate Use?

10 of 40

Speech-to-text conversion is a technology that enables the recognition and understanding of spoken language and its transformation into written text.

Speech-to-Text Conversion

Definition

11 of 40

Speech-to-text conversion is relevant to everyone.
Imagine the following scenario: you’re in Italy and you want to order pizza, but the pizzeria staff only speaks Italian!
We can use the AI capabilities we’ve learned about to help us solve this problem.

In what other scenarios can AI make language more accessible?

12 of 40

Go to Google Translate using this link or by scanning the QR code. (you can access it through both the computer or smartphone device)
Set the source language to English and the target language to Italian.
Click on the microphone button.
If a permission notification pops up, allow microphone access.
Think about your favorite pizza toppings.
Start speaking and watch AI types out what you say!

Let’s experiment speech-to-text conversion

פעילות

Activity

An experimental activity on audio and artificial intelligence, with Google Translate's voice typing.

How will we experiment with voice typing?

Go to Google Translate (available on both computer and smartphone).
Set the source language to English and the target language to Italian.
Click on the microphone button.
If a permission notification pops up, allow microphone access.
Think about your favorite pizza toppings.
Start speaking and watch how the AI types out what you say!

Note: It is recommended to explain to the students that there can be errors, as artificial intelligence is not perfect, and it is important to keep that in mind. Therefore, we cannot rely blindly on AI. Before sending a message transcribed by AI, we must review it to ensure there are no mistakes.

After the activity, clarify to the students that this is an extraordinary capability! “It’s not at all simple to transcribe everything you’re saying. Try to think of yourselves in class, try to write down everything the teacher says and keep up – it’s not an easy task, and AI manages to do this quite well!”

Link to Google Translate

https://translate.google.com/?hl=iw&sl=iw&tl=it&op=translate

13 of 40

Share the order you created using artificial intelligence.
Click on the speaker button to have the AI convert the text into speech.

Share Your Order!

14 of 40

Text-to-speech conversion is a technology that takes written text and generates a voice that reads the text aloud.

How Did We Do It?�By Using Text-To-Speech Conversion.

Definition

15 of 40

What Else Can Be Done With Text And Speech Conversion?

Artificial intelligence opens up amazing possibilities for us.
It’s important to remember that it’s designed to enhance our capabilities and work alongside us.
Think of additional examples where this technology can be used to improve our lives - examples you know from daily life or new ideas you’ve come up with.

דיון

Discussion

Let’s connect what we learned about artificial intelligence to our everyday life. This option opens up many possibilities that can help us in our day to day lives, without us even noticing sometimes.

Note: it’s important to show how AI transcription can make our lives easier, how AI can enhance our capabilities and that by using the technology the right way we can improve ourselves and help our environment.

Ask the students: “What else can be done with text and speech conversion?”

Possible answers:

‘Transcribing lectures or lessons’: To avoid missing important material mentioned in class.
‘Transcribing journalistic interviews’: Helps journalists produce articles faster.
‘Transcribing phone calls’: Useful for keeping track of important conversations at work or home.
‘Helping people with hearing impairments’: Real-time subtitles allow people with hearing difficulties to participate in conversations.
‘I don’t know of other uses’: In this case, provide additional examples and discuss how each can improve daily experiences.

16 of 40

What Else Can Be Done Using It?

Here are some examples of where AI transcription can be used

Voice Assistants

(Google Assistant, Siri, Alexa)

Real-Time Translation

Automatic Subtitles for Videos

Disease Detection

Let’s connect what we learned about artificial intelligence to our everyday life.

Explain to the students that AI can ‘listen’ to speech and convert into text in a quick and accurate way. It opens up many possibilities that can help us in our day to day lives, without us even noticing sometimes.

Here are some examples of AI transcription applications:

Voice Assistants (like Siri, Alexa, Google Assistant): Voice assistants listen to your commands, understand speech, and perform actions like sending messages, setting alarms, or searching for information online.
Google Translate for Real-Time Translation: AI transcribes and translates your speech in real time into another language, enabling you to communicate with people who don’t speak your language, whether abroad or with people from other countries.
Automatic Subtitles for Videos: When watching a video, systems can generate automatic subtitles. This is particularly useful when audio isn’t clear or when the video is in an unfamiliar language.
Disease Detection: By analyzing voice data, it may be possible to monitor patients and detect the development of certain diseases.
Real-Time Subtitles in Virtual Meetings: During Zoom lessons or other platform meetings, you can add real-time subtitles, so everything said in the meeting appears as text - helpful for those with hearing difficulties or when there is background noise.
WhatsApp Voice Notes: Instead of listening to long recordings, you can convert them into text and read what was said. This saves time, especially in noisy environments or when you’re unable to listen.�

17 of 40

Loud And Proud

In order to successfully recognize and transcribe what we say, artificial intelligence needs to understand the voice we produce.

18 of 40

How Do The Words We Say Look As A Sound Wave?

Click on the link to the seeing music website

Allow microphone access and start talking!

You can also upload pre-recorded voices.

Demonstration

To illustrate how your voice "looks like", we will use the Seeing Music website. We will do a class demo, where the teacher will show the students how it’s done. If you’d like, you can let the students explore the website themselves.

Go to the website and click on Start Playing. Allow microphone access and start speaking!�You’ll see how the computer creates a sound wave that matches your voice. If you speak in a high pitch, the wave will rise, and if you speak in a low pitch, the wave will drop. If you’d like, let the students open the website and try it themselves so they can see how the computer processes their voice.

In the Example Voice tab, you can upload a pre-recorded voice or a musical instrument sound and see how the computer interprets the audio.

As an additional explanation to the example from the previous slide, in the Visualization tab, you can select the Oscilloscope option to see how your sound wave looks like (just like in the previous example).

Link to the website:� https://creatability.withgoogle.com/seeing-music/

19 of 40

How Do The Words We Say Look As A Sound Wave?

This is how it looks like:

20 of 40

How Does AI Convert Speech Into Written Text?

The system is training on transcribing the audio files.

2

The system is tested with a new audio file to check if it manages to transcribe correctly what is said.

3

AI learns how to “understand” what we are saying by using context and transcribes what we said.

4

The AI collects many audio files with corresponding transcripts .

1

Explain to the students how artificial intelligence converts speech into written text.

Tell the students that the process of transcribing speech into text is based on three main stages:

Receiving Information (Data Collection): The AI receives audio files with corresponding transcripts (pairs of voice and text).
Training: The system trains on transcribing the language. While we won't go into the technical details of the training process, here is a simplified explanation:�It breaks down the words spoken in the audio file into smaller parts called phonemes. Then, it attempts to write the phoneme it heard, understand which phoneme follows, and ultimately creates text from the spoken words. The text generated by the system is compared to the original transcript, and if there are mistakes, the system repeats the transcription process. This is how the system improves over time.
Testing: At this stage, we check whether the AI can correctly transcribe new, unheard audio into text.

Artificial intelligence uses all the words in the sentence to understand what we are saying. It relies on context to determine whether we said the word "Can" (is able to) or the word “Ken” (A name of a person).

Follow-up question: "Why do you think artificial intelligence needs a large number of audio files with transcripts?"

Possible answers:

"So it can learn to recognize words better."
"I'm not sure." - Explain that if AI sees many examples, it can identify patterns and understand how words are formed in different contexts.

Dive deeper into the subject using this questions: "What can we do if artificial intelligence fails to accurately transcribe a new audio file?"

Possible answers:

"We need to train it with more examples." – Correct! The more examples it trains on, the better it gets.
"Maybe the examples weren’t good enough." – Correct! It’s important that the examples are diverse so the system can handle different types of voices and accents.

21 of 40

Artificial Intelligence Can Resemble Human Intelligence

We practice, make mistakes and learn from our mistakes.

2

We test ourselves by hearing new words and sentences and using the knowledge we gained to see if we understand.

3

4

1

We collect words in a new language by listening to others speak.

We succeed in learning new words in a new language.

The AI trains on transcribing the audio files.

2

The AI is tested with new audio files to check if it can transcribe what was said correctly.

3

4

1

The AI collects many audio files with corresponding transcripts .

The AI learns how to “understand” what we are saying by using context, and then transcribes what was said.

Let’s compare the learning process of AI to the learning process of humans: "The process that artificial intelligence goes through is very similar to how we, as humans, learn new things - like learning a new language."

Receiving Information (Data Collection):

Humans: We gather words in a new language by listening to others and seeing the word written.
Artificial Intelligence: Receives audio files with corresponding transcripts.

Training (Practice):

Humans: We practice speaking, make mistakes, and learn from the corrections we receive.
Artificial Intelligence: Trains by attempting to transcribe audio files and comparing the result to the known transcript. It improves itself based on mistakes and successes.

Testing (Application):

Humans: We encounter new words and sentences and use intuition and context to understand them.
Artificial Intelligence: Like humans, AI uses context to understand what is being said. We test it with new audio files to check if it can transcribe them correctly.

Explain to the students that just as humans learn a language through exposure, practice, and testing, artificial intelligence also improves and becomes more accurate as it trains on diverse examples.

Using these tools, we can create AI systems like the one in Project Relate we saw at the beginning of the lesson - an AI that can help the community.

Note: Humans and AI do not learn in exactly the same way. However, there is a similarity between them, as both humans and AI learn effectively from examples.

22 of 40

What Else Can Be Done Using Artificial Intelligence And Sound?

דיון

Discussion

23 of 40

What Else Can Be Done Using AI And Sound?

Creating songs, melodies, and beats. Try it for yourself!

Now let’s see another use of artificial intelligence and audio - converting a drawing into a melody.

פעילות

Activity

Now, let's present another example of creative use of AI in the field of audio - creating original music using AI-based tools like Suno and Musical Canvas.�We will use Suno and Musical Canvas. These are AI-based tools that allow us to create music. You can choose a style, musical instruments, and even write songs on topics that interest us.

First, we’ll explain about Musical Canvas - through this tool, we’ll explore the unique capabilities of AI and audio by creating musical instruments using AI.

Dive deeper into the subject by asking the students this question: "Why do you think AI is also suitable for creating music?"

Possible answers:

"Because it can learn many music styles." – Correct, AI learns from many examples of songs and musical styles.
"Because it can connect words and melodies." – Exactly, AI can combine verbal and instrumental content.

In the following slides, you’ll find the instructions for each of the two tools. We recommend going over them before the lesson.

Important Notes:

If it is not possible for the students cannot do the activity on their own, do it as a group activity where the teacher guides the students and creates a class song together with them.
Emphasize the importance of privacy and not mentioning personal names.

Links to the websites:

https://suno.com

https://artsandculture.google.com/experiment/8QFo2oQr2uT3pg

24 of 40

Let’s Create A New Melody Using AI With Musical Canvas

To access the instructions sheet for the website, click the link or scan the QR code.

פעילות

Activity

25 of 40

Share With The Class The Melodies You Created Using Artificial Intelligence

פעילות

Activity

26 of 40

Let's Take It Up A Notch:

Create A Song With Suno, Using AI!

To access the website, log in using your e-mail address.
You will have 5 attempts to create a song, so use them wisely.

Demonstration

27 of 40

Go To The Website Suno.Com

Demonstration

28 of 40

Click On ‘create a Song’

Demonstration

29 of 40

Log In To The Website

Demonstration

30 of 40

Choose ‘create’ From The Menu

Demonstration

31 of 40

Now It’s Time To Create Our Song!

Together as a class, choose a subject, for example: a song about your favorite pizza!
What did you choose?
Write the song description in the box
AI will create a song according to the description you gave it.

Demonstration

Now, we’ve reached the stage where we create our song.

Ask the students to write a sentence describing the song they want to create. Use an example, such as: "A song about my beloved dog Lucy and about how I don’t like cream cheese."

Add details about the desired style, for example: "Make the song in a pop style with trumpets and a catchy chorus."

The more details you provide, the more the song will fit your preferences and style.

Write the song description in the box. You can also write in different languages.

Important Points:

Create the song as a group activity, with the teacher guiding the students to collaboratively create the class song. Instructions for using the website are provided in the following slides.
Emphasize the importance of privacy and not including names or personal information - we won’t input personal or identifying data into the AI. AI is a developing tool and uses the information we provide to learn.
To access the website, log in or create an account if you don’t already have one.
You have 5 attempts to create a song, so use them wisely.

32 of 40

Now it’s time to create our song!

click on ‘Create’.

Demonstration

33 of 40

Clicking The ‘custom’ Button Will Allow You To Add More Details

Demonstration

34 of 40

Share With The Class The Song We Produced Using Artificial Intelligence

35 of 40

Creativity In The Age Of Artificial Intelligence

We’ve seen how easy it is to create music using artificial intelligence. With this type of technology, anyone can create a song.

This new technology raises questions: What is the role of artists and creators in the age of artificial intelligence?

Is a musician necessarily someone who knows how to sing, play an instrument, or compose?

דיון

Discussion

36 of 40

What Did We Do?

Creating songs using AI- We typed a song description and chose the musical style to match it.

Creating new music using a combination of drawing and AI: we created new melodies that were influenced by our drawing style.

Suno

Musical Canvas

What Did We Learn?

Artificial intelligence allows us to create music and new sounds even without having musical knowledge.
The connection between humans and technology enables us to be innovative and create in a unique way.

Let’s summarize the activity we just completed.

Explain to the students that in this activity, we explored the creative power of artificial intelligence - another way this technology can be utilized. By using Suno and Musical Canvas, we created new songs and melodies without prior knowledge or musical training. We saw how our descriptions could guide and refine the AI, helping us create a song tailored to our requests.

We also observed that collaboration between humans and AI resulted in new and relevant songs, highlighting the importance of this collaboration. AI is here to enhance our abilities, and this collaboration allows us to achieve new levels of creativity.

Key points to emphasize during the explanation:

Highlight how easy it is to produce professional results using AI.
Make it clear that the process is joint effort: humans provide the instructions and guidance, and AI executes.

Dive deeper into the subject using this question- �Ask the students: "What could you do to improve your song?"

Possible answers:

"Change the song description to add more details." – In response you can say "That’s absolutely right! Adding precise details like style, emotion, or instruments can help the AI better match your expectations."
"Adjust the description to include specific emotions, like happiness or sadness." – You can say: "Exactly! Focusing on the emotion you want to express can enhance the song and make it more accurate."
"I’m not sure." – Explain: "You can try adding or changing details in the song description, like style, instruments, or tempo, and see how it impacts the result."

37 of 40

Where in your daily life can you use the AI technologies we learned about?

How can we use artificial intelligence to help the people around us?

What have you learned about the connection between technology and creativity?

Summary

Ask the students some reflective questions and listen to their thoughts and insights based on what was taught in the lesson.� There are no right or wrong answers here - this is about sharing and reflecting on the material covered.

Where in your daily lives could you use this technology?� Possible answers:

"In conversations with friends from abroad using real-time translation."
"On Zoom, with real-time subtitles to help understand meetings."

How can we use this technology to help people around us?� Possible answers:

"To assist people with hearing impairments by providing live subtitles."
"To create special, personalized songs for anyone who wants one."

What did you learn about the relationship between technology and creativity through using the tools?� Possible answers:

"Technology can help us create things we wouldn’t be able to do on our own."
"Technology gives us new tools that allow us to compose music."
"I’m not sure." – Explain: "This tool shows how technology can transform simple ideas into original sounds. It’s an example of how technology helps us be creative in new ways."

Reflect on the lesson - What have we done so far? Let’s look back on what we did today:

We were introduced to Project Relate, an inspiring project that demonstrates the power of artificial intelligence.
We learned what AI is and how it works.
We explored the concept of "voice" and how AI transcribes speech into text (Speech-to-Text).
We discovered how AI can be used in creative fields, like songwriting.
And we experimented with AI-based tools ourselves!

To summarize: Today we saw how artificial intelligence can expand our capabilities as humans and help us solve diverse problems. It’s important to remember that this technology relies on the collaboration with us, humans, in order to succeed.

38 of 40

The project we saw earlier is only one of many other incredible projects that can help people with disabilities, using AI.
These projects are developed by people right here in Israel.
Let’s meet one of them, Sapir, who works at Google and helped develop such a life changing product, using AI.

Remember Project Relate That Helps People With Disabilities?

39 of 40

40 of 40

Artificial Intelligence Is Here To Enhance Our Capabilities, It Is Both The Present And The Future.

Stay Curious