Lesson 5. AI: Technology, Applications, and Ethics

• Religious education
• Upper Secondary (Key Stage 4)
• Further Education (Key Stage 5)
• BTEC, GCSE

AI

Learning Objectives

• You can explain the definition of artificial intelligence.
• You can give examples of AI applications in daily life.
• You can explain the concept of machine learning.
• You can explain what deep learning is.
• You can explain what robotics is.
• You can list the main advantages of AI.
• You can describe the potential disadvantages and risks of AI, such as ethical dilemmas and biases in algorithms.

What do you know about AI?

How long have we been talking about AI?

1. For about 10 years
2. since 2000
3. Since the 1950s
4. The past 3 years

What is AI?

1. A computer program that can fully understand and perform human tasks independently in all situations
2. A system that can mimic human decisions and reasoning in certain contexts, but has no consciousness or general intelligence
3. Technology that can perform or assist human tasks by learning from data, but is usually specialized in specific tasks
4. A technology that performs advanced calculations and data processing to generate results without actually “understanding” what it is doing

 Machine Learning is a part of AI that focuses on techniques allowing computers to learn based on input data and patterns.

A programmer is needed to intervene and adjust the algorithm.

With deep learning, the algorithms themselves determine whether their decisions are right or wrong.

What is AI (still) not good at?

1. Recognizing images
2. Recognizing emotions
3. Recognizing speech
4. Algorithms

In which country was ChatGPT banned for a month?

1. China
2. Russia
3. North Korea
4. Italy

The Italian privacy authority, Garante, imposed a ban on OpenAI (the developer of ChatGPT) on March 31 from processing personal data of Italian users. According to Garante, OpenAI had not implemented sufficient safeguards to protect users’ privacy. The ban was lifted on April 28.

Machine learning

Machine learning is a way for computers to learn and improve on their own through data and experience, without being specifically programmed for each task.

Email spam filter

Weather and price predictions.

What is Machine Learning?

Machine Learning (ML) is a technology where computers learn from data to get better at a task without being specifically programmed for that task. In other words, the computer learns through experience, just like humans do.

How does Machine Learning work?

In machine learning, a computer is fed large amounts of data. The computer uses this data to find patterns and make decisions. The more data the computer has, the better it can learn and predict.

Example: Suppose you want to create a machine learning system that can recognize cats in photos. You would show the system many photos of cats and non-cats. The system then learns what a cat is by looking at similarities in the cat photos and differences with photos without cats.

There are three main types of machine learning:

Supervised Learning: This is when the computer learns from example data that already has the correct answers. For example, a dataset of images of cats and dogs already labeled as “cat” or “dog.” The computer then learns to recognize the difference.

Unsupervised Learning: Here, the computer gets data without labels and must find patterns on its own. For example, the computer sees different types of fruit without knowing which type is which. It must discover on its own that there are apples, oranges, and bananas.

Reinforcement Learning: The computer learns through rewards and penalties. This works a bit like training a pet. If the computer does something right, it gets a reward; if it does something wrong, it gets no reward or a penalty. This helps it learn what to do and not do. It’s often used in game development.

Applications of Machine Learning:

Machine learning is used in many everyday applications, such as:

Search Engines: Machine learning helps search engines like Google show better results by learning from what users click.

Social Media: Websites like Facebook and Instagram use machine learning to recommend content you might like.

Facial Recognition: Machine learning is used in facial recognition technologies, such as unlocking your phone with your face.

Voice Assistants: Virtual assistants like Siri and Alexa use machine learning to better understand you and respond to your questions.

Machine learning is a way for computers to learn and improve on their own through data and experience, without being specifically programmed for each task. It is widely used in technologies we use every day, from search engines to facial recognition and voice assistants.

Deep learning

They are specifically designed to recognize patterns and structures in large amounts of data.

Deep learning models are a type of algorithm designed to understand and process complex data.

They use networks inspired by the way the human brain works.

Many technologies we use daily are a combination of machine learning and deep learning.

The power of deep learning is that it automatically learns features from data.

What is Deep Learning?

Deep Learning is a part of artificial intelligence (AI) and a subcategory of machine learning (ML). It is a method where computers are trained to recognize patterns and make decisions based on large amounts of data. Deep Learning uses neural networks inspired by the way the human brain works.

How does Deep Learning work?

Deep Learning models use multiple layers of neurons connected to each other, hence the term “deep” in Deep Learning. Each layer learns an abstraction of the input data, allowing the system to recognize increasingly complex patterns.

Here is a step-by-step explanation of how this process works:

Input Layer:

The raw data (e.g., an image, text, or sound) is fed into the network.

Hidden Layers:

These layers perform mathematical operations on the data.

Each layer “detects” specific features, such as edges, shapes, or objects in an image.

The output of each layer becomes the input for the next layer.

Output Layer:

The final layer provides a prediction or classification, e.g., “dog” or “cat” in an image recognition task.

Each neuron in a layer has a weight and a bias, which are adjusted during training. By processing millions of examples, the model learns how to correctly classify or predict input data.

Deep Learning differs from traditional machine learning in how it automatically learns features without human intervention. In classical machine learning, a programmer must manually design “features,” but in Deep Learning, the model learns them on its own. This makes it ideal for complex tasks such as:

Image Recognition (e.g., facial recognition on your phone)

Speech Technology (e.g., Siri or Google Assistant)

Translations (e.g., Google Translate)

Self-Driving Cars (detecting objects like pedestrians and traffic signs)

Medical Imaging Analysis (e.g., detecting tumors in X-rays)

Chatbots

Chatbots: automated programs that simulate human interaction through conversations. They are often used for customer service and providing information.

What are Chatbots?

Chatbots are software programs designed to simulate conversations with humans, usually through text or voice. They use artificial intelligence (AI) and natural language processing (NLP) to understand user input and respond in a way that seems as natural as possible. Chatbots can be programmed for simple tasks, like answering frequently asked questions, or more complex interactions that require understanding context and nuance.

How do Chatbots work?

Chatbots use various technologies such as:

Natural Language Processing (NLP): Enables chatbots to understand human language, including grammar, sentence structure, and sometimes even emotion or intent.

Machine Learning: Allows chatbots to learn from previous interactions and improve at understanding and responding to questions over time.

Rule-based Systems: These chatbots follow predefined rules and scripts. They are limited to performing specific, pre-programmed tasks and are less flexible than AI-based chatbots.

Examples of Chatbots

Siri and Google Assistant:

Although technically digital assistants, they function as chatbots by handling text or voice interactions with users, answering questions like “What time is it?” or “What’s the weather today?”

Duolingo:

The language app uses chatbots to help users practice a new language by simulating conversations to improve language skills.

KLM Royal Dutch Airlines’ BlueBot (BB):

KLM’s chatbot helps passengers book flights, check in, and answer frequently asked questions about their travel. BB communicates via platforms like Facebook Messenger and WhatsApp.

ChatGPT:

ChatGPT is an advanced AI chatbot developed by OpenAI, designed to conduct natural conversations and assist users with tasks like answering questions, writing texts, and generating ideas.

Applications of Chatbots

Chatbots are used in many sectors and situations, including:

Customer Service: Companies use chatbots to quickly and efficiently answer customer questions without human intervention.

E-commerce: Chatbots help customers navigate online stores, recommend products, and complete purchases.

Education: Educational chatbots assist students with homework, explain topics, and provide personalized learning experiences.

Healthcare: Chatbots help patients schedule appointments, provide basic medical information, and manage medication reminders.

Chatbots are becoming increasingly sophisticated as AI technologies develop, enabling them to conduct more complex and natural conversations.

BB Blue Both KLM

Siri and Alexa are not chatbots in the traditional sense, but they do share some similarities with chatbots.

Siri and Alexa exhibit characteristics of a chatbot, such as interacting with users and using natural language processing.

Google, Alexa, Siri

They use AI (artificial intelligence) and various AI technologies to understand speech, respond, and improve at their tasks. Here’s how they use AI:

Speech Recognition (ASR - Automatic Speech Recognition)

Both Siri and Alexa convert your spoken words into text. This process uses machine learning models that recognize patterns in speech. The system gradually learns to distinguish different accents, voice tones, and background noises.

Natural Language Processing (NLP)

After your speech is converted into text, the AI needs to understand what you mean. This is where NLP comes in. The AI system “understands” sentences, identifies keywords, and extracts the intent behind the message. For example:

"Play a song by Coldplay" → The AI recognizes “play” (action) and “song by Coldplay” (context).

Machine Learning (ML)

Both Siri and Alexa get smarter as they receive more data. Every time you interact with them, their models can learn from your preferences, language use, and speech patterns. This is why Siri can give suggestions based on your habits.

Context Awareness and Personal Recommendations

Both AI assistants use context awareness. They remember previous commands to assist you better. For example:

You: “What’s the weather today?”

You later: “And tomorrow?”

The AI understands that “tomorrow” refers to the weather, even if you didn’t say the word explicitly.

Cloud-based AI

While some processing happens locally (e.g., on your smartphone), much of the heavy AI work is done on cloud servers from Apple (for Siri) and Amazon (for Alexa). This cloud AI is more powerful, faster, and continuously updated.

Social media

Facebook, Instagram, TikTok, and YouTube use ML to make recommendations for videos, posts, or ads.

It looks at your previous interactions to show content you are likely to find interesting.

Social Media and AI

Personalized Recommendations (Content Suggestion)

AI decides what content you see in your feed on platforms like Instagram, TikTok, Facebook, and YouTube. This uses recommender systems that analyze your interests, behavior, and interactions.

AI looks at which posts you like, share, or comment on.

Algorithms predict which posts you are likely to find interesting.

This keeps you on the platform longer, which benefits social media advertising revenue.

Image and Video Recognition

AI can automatically analyze and identify what is in photos and videos. This helps platforms like Instagram and Facebook to:

Suggest automatic tags for photos (e.g., facial recognition).

Detect inappropriate content, such as violence, hate speech, or nudity.

Enable visual search features (like Google Lens).

Content Moderation (Automatic Detection of Harmful Content)

AI helps platforms identify and remove harmful, violent, or inappropriate content. Used by Facebook, Twitter, TikTok, and YouTube.

NLP (Natural Language Processing) scans text for hate speech, bullying, or threats.

Image recognition detects nudity, weapons, or violence in images and videos.

Example: YouTube automatically removes copyright-infringing videos by comparing them with files in the “Content ID” database.

Personalized Ads (Ad Targeting)

AI helps platforms like Facebook, Instagram, and LinkedIn target ads to users.

AI analyzes your behavior (likes, clicks, purchases) and predicts which ads you may find interesting.

Advertisers can reach specific audiences based on interests, age, location, etc.

Chatbots and Customer Service

AI-powered chatbots provide 24/7 support on platforms like Facebook Messenger.

NLP helps chatbots understand customer questions.

The chatbot can respond automatically or escalate to a human agent.

Sentiment Analysis (Opinion Analysis)

AI can analyze the tone of posts (positive, negative, or neutral). Brands use this to monitor their reputation.

NLP detects words and emotions in posts.

Companies can respond quickly if many negative posts appear about their brand.

Bots and Fake Account Detection

AI helps platforms detect spam, fake accounts, and bots.

It looks for irregular behavior, like creating many accounts from the same IP address.

It analyzes user activity (e.g., posting excessively or following hundreds of accounts in one day) to flag suspicious accounts.

Example: Twitter removes millions of fake accounts daily that spread spam or misinformation.

Creative Tools (AR Filters and Effects)

AI allows users to use Augmented Reality (AR) filters in apps like Snapchat, Instagram, and TikTok.

Machine learning detects faces and movements.

Effects (like face filters) move with the user in real time.

Predictive Analytics (Trend Prediction)

AI predicts which trends will become popular based on user behavior.

It analyzes which topics (hashtags, trends) suddenly gain popularity.

Platforms can highlight trending topics or show users posts about current events.

Example: Twitter’s “Trending” section is determined by AI models tracking spikes in hashtag activity.

Data Collection and User Profiles

AI helps platforms build detailed user profiles.

It collects data on your search history, likes, comments, location, and interactions.

This data is used to deliver targeted ads and personalize your experience.

Example: If you frequently view sports content, social media may show you more sports-related posts.

Robotics

Robotics: The design, construction, and use of robots. These robots can be autonomous or remotely controlled.

Robotics is about creating and using robots to help humans with various tasks.

Robotics

Robotics is the field concerned with designing, building, and using robots.

Robots are machines that can be programmed to perform tasks normally done by humans. They can be used for a wide range of purposes, such as building cars, cleaning up waste, performing surgeries, or exploring other planets. Robots vary in shape and size, from simple household robots to advanced industrial and scientific robots.

Examples:

Industrial Robots:

Welding Robots: Used in factories to weld cars and other machinery. They are very precise and can perform the same work repeatedly without getting tired.

Production Arms: Robots with arms that assemble, sort, or package products in factories. Commonly used in the automotive and electronics industries.

Household Robots:

Vacuum Robots: Robots like the Roomba can automatically navigate a home to vacuum floors without human operation.

Lawn Mowing Robots: These robots mow the lawn automatically and return to their charging station when finished.

Medical Robots:

Surgical Robots: Robots like the Da Vinci robot assist surgeons in performing complex operations with high precision, which is difficult with human hands alone.

Rehabilitation Robots: Robots that help people recover after injury or surgery by assisting with physical exercises.

Research and Exploration Robots:

Mars Rovers: Robots like the Mars rover "Curiosity" are used to explore other planets. They can take photos, collect samples, and send data back to Earth.

Underwater Drones: Robots used to explore the ocean floor and inspect pipelines or shipwrecks.

Service Robots:

Restaurant Robots: In some restaurants, robots take orders, serve food, or clear tables.

In short, robotics is about creating and using robots to help humans with a variety of tasks.

Cleaning Robots

Health

Managing the health of a population (predicting epidemics)

Supporting clinical decision-making (detecting brain tumors/eye disorders)

Which statement is correct?

1. With Machine Learning, humans don’t have to do anything
2. More data = better results
3. Machine learning is not yet applicable in education
4. Machine learning has had its day

Machine Learning is a part of AI that focuses on techniques enabling computers to learn based on input data and patterns.

A programmer is needed to intervene and adjust the algorithm.

With deep learning, the algorithms themselves determine whether their decisions are right or wrong.

Advantages of AI

Some advantages of AI

Processing large amounts of data

Accuracy and precision

Efficiency and productivity

Innovation and new opportunities

Advantages of AI

1. Increased Productivity

Automation of repetitive tasks: AI can automate routine and time-consuming tasks, such as data processing, data entry, and inventory management. This reduces human workload and frees people for more complex tasks that require creativity and decision-making.

Faster decision-making: By analyzing large amounts of data in real time, AI can quickly generate insights and recommendations, helping businesses and organizations make faster and more accurate decisions, improving efficiency and productivity.

Smart planning and logistics: AI systems can use advanced planning algorithms to optimize production processes, manage inventory better, and streamline logistics operations, reducing waste and increasing efficiency.

2. Better Healthcare

Advanced diagnosis and treatment: AI can identify patterns in medical data that doctors might overlook, leading to earlier and more accurate diagnoses. AI algorithms can also suggest personalized treatment options based on a patient’s specific characteristics.

Remote care: AI enables healthcare providers to monitor patients remotely via wearables and other technologies, which is particularly useful for elderly patients or those with chronic conditions, ensuring continuous monitoring and timely intervention.

Efficiency in administrative tasks: AI can streamline administrative processes such as scheduling appointments, managing patient records, and processing insurance claims, giving healthcare providers more time to focus on patient care.

3. More Efficient Energy Management

Smart grids: AI can manage energy networks, predict supply and demand, and distribute energy efficiently, reducing losses and improving reliability.

Optimizing energy use: In commercial and industrial buildings, AI can optimize energy consumption by analyzing patterns and making real-time adjustments, like turning off lights or adjusting heating and cooling.

Supporting renewable energy: AI helps integrate renewable energy sources like wind and solar into the energy grid by analyzing weather forecasts and optimizing energy storage and distribution, making sustainable energy use more efficient and reliable.

4. Enhancing Human Creativity

Inspiration and new ideas: AI can inspire artists, writers, musicians, and designers by suggesting new ideas and concepts. By analyzing patterns and styles, AI can propose combinations and variations that humans might not think of.

Generative design: In architecture and product design, AI can generate thousands of design options that meet specific criteria, allowing designers to experiment more and reach innovative solutions faster.

5. Automation of Creative Tasks

Content creation: AI tools can help create music, write texts, and produce visual art, lowering the barrier for people to participate in creative processes, even without traditional training or skills.

Video and image editing: AI can automate complex image and video editing tasks, such as color correction, object removal, or generating special effects, giving creatives more time to focus on the artistic aspect.

6. Personalization and Experience

Custom art and media: AI can create personalized artwork, music, or stories tailored to an individual’s preferences, increasing engagement and making the creative experience more meaningful.

Interactive and adaptive experiences: In gaming and entertainment, AI can create dynamic, responsive environments that adapt to user actions, resulting in more immersive and engaging experiences.

7. Discovering New Styles and Techniques

Style transfer and artistic exploration: AI can help discover new artistic styles by combining existing ones or experimenting with new techniques, allowing artists to explore unconventional methods.

Innovation in music and sound: By analyzing patterns in music, AI can create new music styles or refresh existing genres. It can also generate unique sound effects and compositions beyond traditional methods.

8. Human-Machine Collaboration

Co-creation: AI can act as a creative partner, generating ideas, offering suggestions, and collaborating with humans to create new works, fostering a new form of creativity where humans and machines complement each other.

Feedback and improvement: AI can provide real-time feedback on creative works, such as paintings, music, or writing, suggesting improvements based on analysis of aesthetics, composition, or style.

Disadvantages of AI

Search for the disadvantages of AI yourself using AI. Look at them carefully. Check if the answers are correct. Create your own top 3.

Disadvantages of AI 

Employment and Automation

1

Bias and Discrimination

2

Privacy Concerns

3

Ethical Considerations

5

Dependence on Technology

4

Costs

6

Complexity and Lack of Transparency

10

Security Risks

7

Environmental Impact

9

Social Isolation

8

Disadvantages of AI

Employment and Automation: AI and automation can lead to job losses, especially in sectors with repetitive tasks, which may increase economic inequality and social tensions.

Bias and Discrimination: AI systems can contain biases, particularly if trained on biased data. This can result in unfair or discriminatory decisions in areas like hiring, justice, and credit scoring. AI may reflect the beliefs, political preferences, ethical choices, or religious views of its creators, which can make it less neutral.

Privacy Concerns: AI often collects large amounts of personal data, raising concerns about privacy and data security, especially if the data is misused.

Dependence on Technology: Heavy reliance on AI may cause humans to lose certain skills and become overly dependent on technology, which can be problematic if systems fail.

Ethical Considerations: Questions arise about responsibility for decisions made by AI. Who is accountable if an AI system makes a mistake or causes harm?

Costs: Developing and implementing AI technology can be expensive, making it difficult for smaller companies or organizations with limited resources.

Complexity and Lack of Transparency: Many AI models, especially deep learning models, are hard to understand and interpret, which can cause distrust and uncertainty about decision-making processes.

Security Risks: AI systems can be vulnerable to hacking or misuse, potentially leading to severe consequences in critical areas like healthcare or infrastructure.

Environmental Impact: Training complex AI models requires significant computing power, increasing energy consumption and ecological footprint.

Social Isolation: Increased interaction with AI systems may reduce human-to-human contact, potentially leading to social isolation and mental health issues.

Which disadvantage do you think is the most important?

When AI makes a mistake, there are several important considerations regarding responsibility and liability:

Developers and Manufacturers: Generally, the creators or manufacturers of AI technology are responsible for its functioning. If an error occurs due to negligence in design, training, or testing, they may face legal liability.

Users: Users of AI systems can also be held accountable, especially if they misuse the technology or fail to follow the guidelines provided by the developers. This is particularly relevant when users influence outcomes, such as in decision-making algorithms.

Organizations: Organizations deploying AI may be liable for the consequences of its use, especially in sectors like healthcare, finance, and transportation, where errors can have serious impacts.

Regulations and Legislation: Legal and ethical frameworks for AI are still developing. Countries and regions are creating laws and guidelines to determine liability when AI causes harm, including discussions about whether AI should have a legal status.

Ethical Considerations: Ethical questions arise regarding decisions made by AI, especially in sensitive areas such as healthcare or justice. Determining moral responsibility for AI-driven outcomes is a key issue.

Insurance: Insurance companies can also play a role, providing coverage for damages caused by AI systems, helping assign financial responsibility in case of errors.

Complexity of AI Systems: Many AI systems, particularly those using machine learning, are complex and hard to understand. This makes it difficult to pinpoint exactly where and why a mistake occurred, complicating responsibility assignment.

Transparency and Explainability: Clear explanations of how AI systems work are crucial. Transparency helps understand errors and properly assign responsibility.

All these factors make it challenging to determine who is responsible when AI goes wrong. This remains an ongoing topic of discussion in both legal and ethical contexts.

Who is responsible if AI makes a mistake?

When AI goes wrong, there are several important considerations regarding responsibility and liability:

Developers and Manufacturers: Usually, the creators or manufacturers of AI technology are responsible for its operation. If an error occurs due to negligence in design, training, or testing, they may face legal liability.

Users: Users of AI systems can also be held accountable, especially if they misuse the technology or do not follow the guidelines provided by the developers. This is particularly relevant when users influence outcomes, such as in decision-making algorithms.

Organizations: Organizations deploying AI can be held liable for the consequences of its use. This is especially important in sectors like healthcare, finance, and transportation, where mistakes can have serious impacts.

Regulations and Legislation: Legal and ethical frameworks for AI are still developing. Different countries and regions are creating laws and guidelines to determine liability when AI causes harm, including discussions about whether AI should have a legal status.

Ethical Considerations: Ethical questions arise about decisions made by AI, especially in sensitive areas like healthcare or justice. Determining moral responsibility for AI-driven outcomes is a key issue.

Insurance: Insurance companies may play a role by covering damages caused by AI systems, helping assign financial responsibility if a mistake occurs.

Complexity of AI Systems: Many AI systems, particularly those based on machine learning, are complex and difficult to understand. This makes it challenging to determine exactly where and why an error happened, complicating responsibility assignment.

Transparency and Explainability: Transparency in how AI systems operate and the ability to explain their decisions are crucial. This helps understand errors and assign responsibility.

All these factors make it difficult to clearly determine who is responsible when AI makes a mistake. This is an ongoing topic of discussion in both legal and ethical contexts.

What is the risk of incorrect or limited data in AI?

1. This can lead to unclear or incomprehensible AI systems
2. This can lead to job losses
3. This prevents AI systems from being vulnerable to attacks
4. This can lead to bias and discrimination

 Machine Learning is a part of AI that focuses on techniques allowing computers to learn from input data and patterns.

A programmer is needed to intervene and adjust the algorithm.

In deep learning, the algorithms themselves determine whether their decisions are right or wrong.

What is the importance of privacy and data protection in AI?

1. It is important to strive for fair and unbiased AI
2. It is important to protect users’ privacy
3. It is important that users understand how AI systems work
4. It is important to establish clear responsibilities and liabilities

 Machine Learning is a part of AI that focuses on techniques enabling computers to learn from input data and patterns.

A programmer is needed to intervene and adjust the algorithm.

In deep learning, the algorithms themselves decide whether their decisions are right or wrong.

How could AI be used in education?

1. Personalized Learning Experience

Adaptive Learning: AI can adjust learning programs to the individual needs of students, considering their progress, strengths, and weaknesses.

Intelligent Tutors: Virtual tutors can provide extra support by answering questions and giving explanations tailored to a student’s learning style.

2. Improved Feedback

Automatic Grading: AI can grade assignments and tests automatically, saving teachers time and giving students immediate feedback.

Analytics Tools: AI can analyze student performance and identify areas where extra help is needed, allowing teachers to provide targeted support.

3. Administrative Efficiency

Time Management: AI can help plan lessons, manage schedules, and optimize resource use.

Learning Data Analysis: AI can process large amounts of data to identify trends in student performance and improve policies and strategies.

4. Accessibility and Inclusion

Support for Diversity: AI can provide tools for students with special needs, like speech recognition and translation software, making learning more accessible.

Language Support: AI translators can help bridge language barriers for non-native speakers.

5. Engagement and Motivation

Gamification: AI can help create educational games that make learning fun and increase student motivation.

Interactive Learning Environments: AI-powered virtual and augmented reality experiences can engage students in hands-on learning and explain complex concepts visually.

6. Preparation for the Future

Essential Skills: AI can prepare students for the future by exposing them to technologies and skills relevant to the job market.

Career Guidance: AI-driven platforms can help students explore career opportunities based on their interests and abilities.

Source: How China Is Using Artificial Intelligence in Classrooms WSJ (YouTube)

How do you think AI will affect your future?

 Think about possible career opportunities and personal interactions with AI technology.