Artificial Intelligence: Powering the Future

Artificial Intelligence
(AI): An Introduction
Prepared by: Kumari Shivangi

 Artificial Intelligence (AI) is the science of
making machines think, act, and learn like
humans.
 It focuses on enabling computers to perform
tasks that usually require human intelligence.
Examples include:
 Speech recognition (Alexa, Siri)
 Image recognition (Face ID)
 Decision-making systems (Google Maps,
medical diagnosis AI)
What is AI?

 Automation – perform repetitive tasks without
human help.
 Decision-Making – AI can evaluate data and
suggest or take actions.
 Learning Capability – improves performance
through experience.
 Adaptability – adjusts to changing conditions.
 Mimics Human Intelligence – reasoning,
planning, problem-solving.
Key Features of AI

 1950s Birth of AI; Alan Turing proposes “Can
→
machines think?” Turing Test.
→
 1960s–1970s Early AI programs; basic problem-
→
solving & symbolic systems.
 1980s Expert systems for medical diagnosis and
→
engineering.
 1990s AI beats humans in games (IBM’s Deep Blue
→
beats Garry Kasparov in chess).
 2000s Rise of Machine Learning (data-driven AI).
→
 2010s onwards Deep Learning + Big Data + GPUs
→ →
self-driving cars, healthcare AI, real-world adoption.
Evolution of AI (Timeline)

 Healthcare disease detection, robotic surgery,
→
personalized medicine.
 Finance fraud detection, risk assessment,
→
algorithmic trading.
 Retail personalized recommendations, chatbots,
→
supply chain optimization.
 Education adaptive learning systems, smart
→
tutoring.
 Transportation autonomous vehicles, route
→
optimization.
 Agriculture crop monitoring, yield prediction.
→
Scope of AI

 Basic AI (Narrow AI): The majority of AI nowadays is
"narrow," intended for particular tasks. Within the
Scope of Artificial Intelligence, examples are
language interpreters and recommendation systems.
 Advanced AI (General AI): This AI, called AGI, could
do a variety of duties much like a person. The Scope
of Artificial Intelligence still includes this intriguing
theoretical area.
 Future AI (Super AI): Super AI, which surpasses
human intellect, is still a theoretical concept with
undetermined potential. Although far off, this future
level of AI has potential.

 Reactive Machines: These AIs lack memory and learning
capabilities, but they are capable of making judgments
based on real-time input. This Limited Scope of Artificial
Intelligence is best shown by chess-playing AI.
 Limited Memory AI: Similar to how self-driving vehicles
use recent knowledge to traverse highways more safely,
this AI temporarily recalls historical data to enhance
present judgments.
 Self-Aware AI: Self-aware AI is a theoretical concept that
would comprehend intents and emotions and interact
with humans on a human-like level, albeit this is probably
a long way off.

 Performs a specific task extremely well but
cannot generalize.
 Present in most real-world applications today.
 Examples:
◦ Email spam filters
◦ Google Translate
◦ Netflix or Amazon recommendations
◦ Voice assistants like Alexa & Siri
Narrow AI

 Represents machines with human-level
intelligence.
 Can perform multiple different tasks and
adapt like a human.
 Currently does not exist – still a research goal.
 If achieved, it could revolutionize industries but
also raises ethical concerns.
General AI

 Future vision where AI becomes smarter than
humans in every aspect.
 Could exhibit creativity, reasoning, and
emotional intelligence beyond human
capacity.
 Theories suggest it could either:
◦ Solve world problems (disease, climate change) OR
◦ Pose existential risks if misused.
Super AI

 Basic form of AI; no memory, no learning.
 Works only with current real-time input.
 Cannot use past experiences to improve
decisions.
 Example: IBM’s Deep Blue (chess-playing AI).
Reactive Machines

 Can use past data temporarily to make better
decisions.
 Learns from recent experiences but memory is
not permanent.
 Widely used in present-day AI systems.
 Example: Self-driving cars (analyzing recent
traffic data to adjust speed, brakes, or turns).
Limited Memory AI

 Future stage of AI with consciousness,
emotions, and self-awareness.
 Could understand human feelings and
intentions.
 Capable of making independent decisions like
humans.
 Still a research concept — does not exist yet.
Self-Aware AI

 Data: AI needs huge amounts of quality data to
learn.
 Algorithms: Rules/models that process data
(Machine Learning, Deep Learning).
 Computing Power: GPUs, TPUs, cloud systems
to handle big data.
 Human Input: Feedback, training, and
supervision to improve models.
Components of AI Systems

 Definition: A subset of AI where machines
learn patterns from data.
 How it works: Provide input train model
→ →
make predictions.
Types of ML:
 Supervised Learning: Learn from labeled data.
 Unsupervised Learning: Find hidden patterns
in unlabeled data.
 Reinforcement Learning: Learn by trial and
error using rewards/punishments.
Machine Learning in AI

 Subset of ML inspired by the human brain’s
neural networks.
 Uses layers of artificial neurons for decision-
making.
Applications:
 Face recognition (Facebook tagging)
 Speech-to-text (Google Assistant)
 Autonomous vehicles (Tesla, Waymo)
 Medical imaging (tumor detection in X-rays)
Deep Learning in AI

 Artificial Intelligence (AI): The broad concept of
machines simulating intelligence.
 Machine Learning (ML): A subset of AI; machines
learn from data.
 Deep Learning (DL): A subset of ML; uses deep
neural networks for complex tasks.
🔹 Example:
 AI Self-driving car as a whole system.
→
 ML Algorithm that learns to detect pedestrians.
→
 DL Neural network that recognizes objects in
→
camera images.
AI vs ML vs DL

 Fairness: AI must not discriminate (e.g., job
hiring systems).
 Privacy: Protect sensitive user data.
 Transparency: AI should explain its decisions.
 Responsibility: Who is accountable when AI
makes mistakes?
Ethics in AI

 Cause: Biased or incomplete training data.
Examples:
 Facial recognition struggles with darker skin
tones.
 Hiring algorithms biased against women.
Solutions:
 Use diverse and representative datasets.
 Test for fairness before deployment.
 Build explainable and auditable AI systems.
Bias in AI

 Human-AI collaboration for problem-solving.
 Advanced healthcare systems (AI doctors,
precision medicine).
 AI for climate change solutions.
 Smarter automation in industries.
 Ongoing debate: benefits vs risks of AI
dominance.
Future of AI

 AI = powerful technology shaping the future.
 Narrow AI dominates today, General & Super AI
remain futuristic.
 Must address ethics, fairness, and bias.
 Balanced approach = harnessing AI benefits
while minimizing risks.
Conclusion

Artificial Intelligence: Powering the Future

More Related Content

Similar to Artificial Intelligence: Powering the Future

More from KumariShivangi5

Recently uploaded

Artificial Intelligence: Powering the Future