Agentic AI: 4 Reasons Why It’s the Next Big Thing in AI

What is Agentic AI? 4 reasons why it is the next big thing in AI research

Agentic AI is a new way to make AI smarter. It lets AI make choices, adapt, and learn from mistakes. This makes it a big deal in AI research. To learn more, visit eventasia.net.

Agentic AI uses machine learning and goal-setting to solve problems better. It can give you experiences and solutions that are just right for you. This is a big step forward in AI, letting systems think and act on their own.

Agentic AI is more independent, needing less human help. It can understand and talk in natural language. This makes it better at working with humans. Once trained, it can be used in many areas, showing it’s scalable.

Agentic AI is exciting because of its many uses. It can help in healthcare and finance, among other fields. This makes it a focus for those looking into AI’s future.

As agentic AI grows, it will change many industries. It will make tasks easier and improve customer service. It will also help humans and AI work better together, without replacing us.

Agentic AI learns and gets better over time. This makes it a game-changer for AI. Its many uses make it a fascinating area to study.

Key Takeaways

  • Agentic AI enables autonomous decision-making and adaptability, making it a significant advancement in AI technology.
  • Agentic AI combines machine learning capabilities with goal-oriented behavior for improved problem-solving, which is a key aspect of what is agentic ai.
  • Agentic AI can process natural language and demonstrate reasoning in decision-making, making it a critical factor in agentic ai research.
  • Agentic AI has the ability to revolutionize various industries, from healthcare to finance, making it the next big thing in AI research.
  • Agentic AI enables increased efficiency and productivity by handling complex tasks and promoting strategic human-AI collaboration, which is a key benefit of agentic ai.
  • Agentic AI is capable of learning from its environment and refining its understanding over time, making it a key area of focus for those interested in agentic ai and the next big thing in AI research.

Understanding the Evolution of AI: From Traditional to Agentic Systems

The evolution of AI technology has brought about advanced artificial intelligence. This change has transformed how we solve problems and make decisions. Agentic AI, a new type, is set to revolutionize many industries.

Traditional AI systems can’t work on their own or make decisions based on current data. Agentic AI, on the other hand, can adapt and make smart choices. This makes it more efficient and effective. Agentic AI’s benefits include being able to work alone, adapt, and improve based on feedback.

Some key differences between traditional and agentic AI include:

  • Autonomy: Agentic AI agents can work alone, making operations more efficient and responsive.
  • Adaptation: Agentic AI agents can change and improve their strategies based on feedback, staying relevant.
  • Action: Agentic AI agents can turn insights into real actions, influencing the operational landscape.

Big tech companies like Salesforce and HubSpot are investing a lot in agentic AI. This has greatly improved customer experience and made operations more efficient. For instance, Salesforce uses agentic AI to offer personalized, real-time customer support.

CompanyApplication of Agentic AIBenefits
SalesforcePersonalized customer supportEnhanced customer experience
HubSpotAutomated email marketing and customer serviceIncreased efficiency and productivity

What is Agentic AI? 4 Reasons Why It Is the Next Big Thing in AI Research

Agentic AI is a new way to make AI smarter. It lets AI make choices on its own, change plans, and keep learning. Studies show agentic ai research is very promising. It’s becoming the next big thing in ai research.

Agentic AI can act alone, making choices based on what it sees and learns. This makes it very useful.

Here are some reasons why agentic AI is so exciting:

  • Autonomous decision-making: Agentic AI can decide things by itself, without always needing a human to tell it what to do.
  • Real-time adaptability: It can change its plans quickly when new information comes in. This makes it more efficient and effective.
  • Iterative learning: Agentic AI keeps getting better by learning from its mistakes and feedback. This leads to better performance over time.
  • Goal-oriented behavior: It works towards specific goals, not just doing one thing at a time. This makes it more effective in reaching its goals.

Agentic AI can change many industries because it can work alone, adapt quickly, and learn from its mistakes. It’s a big deal in what is agentic ai and agentic ai research. As AI keeps getting smarter, agentic AI will be a big part of its future.

The Revolutionary Architecture of Agentic AI Systems

Agentic AI systems work on their own, needing little human help. They are perfect for tasks that need constant watching or quick choices. This is thanks to advanced artificial intelligence and ai technology. They learn from their actions and decide based on current data.

The benefits of agentic ai are many. They can offer custom experiences and solutions. They learn from how users interact with them, helping businesses a lot. For instance, AI agents can handle supply chains, manage stock, and predict demand. This makes operations more efficient and cuts costs.

Agentic AI systems have key parts like decision-making tools and learning abilities. These parts let them work alone and make choices based on new data. Their decision-making tools focus on being independent, flexible, and always getting better. This makes them better than old AI models.

Core Components of Agentic AI

  • Decision-making frameworks
  • Advanced learning capabilities
  • Autonomy and adaptability

With these main parts, agentic AI systems help businesses grow and stay ahead. They make sales and marketing smoother, using automated and custom campaigns.

The Power of Autonomous Decision-Making in Agentic AI

Agentic AI lets it make choices on its own, using current data. This is a big step up from old AI, which only did one thing. Now, agentic AI works towards clear goals, marking it as the next big thing in ai research.

Agentic AI brings many benefits, such as:

  • It can decide on its own, reacting fast to new info or changes
  • It keeps getting better with feedback
  • It can change plans quickly in many fields, helping companies stay strong in uncertain times

Big names like Amazon and UPS are using agentic AI. Amazon changes prices based on demand, and UPS adjusts delivery routes. Hospitals also use it to change patient plans as needed. These examples show how agentic AI can change many fields.

agentic ai

As agentic ai research grows, we’ll see more cool uses. Agentic AI can act alone and decide fast, based on new data. It’s set to change how we solve big problems and make choices.

IndustryApplication of Agentic AI
HealthcareAdjusting patient treatment schedules in real-time
FinanceDynamically adjusting product pricing based on market demand
TransportationRerouting delivery trucks in real-time to avoid traffic congestion

Transformative Applications of Agentic AI Across Industries

Agentic AI is changing many industries, making them more efficient and productive. Thanks to fast progress in AI research, new AI tech is being used in healthcare, finance, and transportation.

Agentic AI has brought big improvements to these fields. For example, more than half of service professionals report noticeable improvements in customer support thanks to AI.

Healthcare and Medical Research

In healthcare, agentic AI helps improve supply chains. This means businesses can handle challenges faster and better. It also makes customers happier.

Financial Services and Trading

In finance, AI is making trading smoother and faster. It cuts down on the need for human help. This boosts efficiency and saves time, with AI-powered content tools saving marketers an average of around three hours per piece.

Manufacturing and Supply Chain

In manufacturing, AI optimizes supply chains. This helps businesses tackle challenges better and faster. It also lowers costs and boosts productivity.

Environmental Protection and Sustainability

In environmental protection, AI helps use resources better. It cuts down waste and supports sustainability. This is key for companies aiming to reduce their environmental footprint.

IndustryApplication of Agentic AIBenefits
HealthcareOptimizing supply chainsReduced response times, increased customer satisfaction
FinanceOptimizing trading processesIncreased efficiency, productivity
ManufacturingOptimizing supply chainsReduced operational costs, enhanced productivity
Environmental ProtectionOptimizing resource usageReduced environmental impact, promoted sustainability

Breaking Barriers: How Agentic AI Overcomes Traditional AI Limitations

Agentic AI is changing the game in artificial intelligence. It allows for self-directed actions, flexibility, and continuous learning. What is agentic ai is key in agentic ai research as it tackles old AI hurdles. Gartner says by 2024, 15% of daily work decisions will be made by Agentic AI, a big jump from 0% now.

The next big thing in ai research is Agentic AI. It’s set to make 15% of work decisions on its own by 2028. Agentic AI is used in self-driving cars, robots, virtual assistants, and even in video games. It’s creating a virtual workforce to help and support human tasks.

Here are some key statistics that highlight the growth of Agentic AI:

  • By 2028, 33% of enterprise software applications are expected to include Agentic AI, up from less than 1% in 2024.
  • Deloitte predicts that in 2025, 25% of companies using generative AI will launch agentic AI pilots or proofs of concept, increasing to 50% by 2027.
  • Investors have invested over $2 billion in agentic AI startups in the last two years, with a focus on companies targeting the enterprise market.
agentic ai

As Agentic AI grows, it will impact many sectors like healthcare, finance, and transport. It can work on its own, make decisions with current data, and learn from its actions. Agentic AI is set to be a major player in artificial intelligence.

YearAgentic AI Adoption
2024Less than 1%
202525%
202750%
202833%

The Role of Agentic AI in Shaping Future Technologies

As ai technology grows, advanced artificial intelligence is spreading across many fields. Agentic ai stands out because it can work on its own and make choices based on current data.

Agentic ai is set to change how we use IoT and smart systems. It will help make systems more efficient and cut down on mistakes. For example, agentic ai can plan and do complex tasks on its own. This makes it perfect for industries like manufacturing and logistics.

Some key traits of agentic ai are:

  • Autonomous decision-making
  • Real-time data analysis
  • Adaptability to new information
  • Interaction with other ai agents and systems

These traits make agentic ai very appealing to industries wanting to automate and improve their work.

The future of agentic ai will be shaped by its connection with IoT and smart systems. As ai keeps getting better, we’ll see more advanced uses of agentic ai in different fields.

IndustryPotential Application
HealthcarePatient data analysis and personalized treatment plans
FinanceAutomated trading and investment decisions
ManufacturingPredictive maintenance and quality control

In summary, agentic ai will play a big role in shaping future tech. Its ability to automate and improve various industries is huge. As ai keeps advancing, we’ll see even more advanced uses of agentic ai.

Ethical Considerations and Responsible Development of Agentic AI

As agentic ai grows, it’s key to think about ethics and how to use it right. Agentic ai research could make the AI market huge, hitting $407 billion by 2027.

But, making agentic ai raises big questions about fairness, openness, and who’s to blame. We must make sure agentic ai systems are clear and open. This means using tests and talking about the tough parts like how it changes and affects others.

Some big hurdles in agentic ai include:

  • It can act in ways we can’t predict and might not want.
  • It’s hard for us to understand and might be unsafe.
  • It’s hard to make sure it does what we want and is fair.

By tackling these issues and focusing on doing it right, we can make agentic ai help us improve things. This way, we can build a brighter future for everyone.

Preparing for an Agentic AI-Driven Future: What You Need to Know

We’re heading towards a future where AI is in charge. It’s key to know how agentic AI can change many fields. AI research is growing fast, promising big leaps in technology.

Agentic AI can act on its own, make choices with current data, and learn from its actions. To get ready for this future, we need to learn and grow. Focus on these areas:

  • Learn the basics of agentic AI and its uses
  • Get good at AI research and using it
  • Keep up with the newest AI tech

Investment in AI worldwide is set to jump by 2027. Growth is expected in healthcare and telecom. With the right skills, you can lead in this AI revolution.

By embracing agentic AI and keeping up with AI research, we can reach AI’s full power. This will make a better future for everyone.

IndustryForecasted Growth
Healthcare50% annual growth rate
Telecommunications$3.7B to $9.3B in the next four years
Financial ServicesMore than double by 2027

Conclusion: Embracing the Agentic AI Revolution

Agentic AI is changing how we solve problems, make decisions, and innovate. This agentic AI technology is more than just the next step in AI research. It’s a big change in how we use artificial intelligence to face tough challenges.

In healthcare, it can improve patient care. In business, it can make supply chains better and reduce financial risks. Companies that use what is agentic AI will get ahead. They’ll work better, make fewer mistakes, and be more efficient.

As technology gets more connected, agentic AI will play a bigger role. Knowing how agentic AI can change things helps us and our companies grow. The agentic AI revolution is here. It’s time to join in.

FAQ

What is Agentic AI?

Agentic AI is a new AI technology that can work on its own. It makes choices based on current data and learns from its actions. This technology is a big step forward in AI and could change many industries.

What are the key features of Agentic AI?

Agentic AI can work alone, make choices with current data, and learn from its actions. These abilities make it different from old AI systems. It’s very flexible and useful in many areas.

Why is Agentic AI considered the next big thing in AI research?

Agentic AI is seen as a major breakthrough in AI research. It can do things that old AI systems can’t. Its ability to make decisions, process data, and learn makes it very powerful for many uses.

How does the architecture of Agentic AI systems differ from traditional AI?

Agentic AI systems have a new design. They have key parts, advanced decision-making, and learning abilities. These help Agentic AI work alone, make choices with current data, and get better over time.

What are the possible uses of Agentic AI?

Agentic AI can be used in many fields like healthcare, finance, transportation, and protecting the environment. Its ability to make decisions and process data makes it great for solving big problems and driving new ideas.

How does Agentic AI overcome the limitations of traditional AI?

Agentic AI works around old AI’s limits by working alone, making choices with current data, and learning from its actions. This lets Agentic AI adapt and solve complex problems better than old AI.

What is the role of Agentic AI in shaping future technologies?

Agentic AI will be key in creating new technologies like IoT, smart systems, and working with humans and AI. Its advanced abilities and flexibility make it essential for next-generation tech that will bring new changes and ideas.

What are the ethical considerations in the development of Agentic AI?

Making Agentic AI must focus on ethics like being open, accountable, and fair. It’s important to develop Agentic AI responsibly. This ensures it’s safe and helpful in many areas.

How can individuals and organizations prepare for an Agentic AI-driven future?

To get ready for an Agentic AI future, people and groups need to learn new skills, keep up with trends, and look into investing in this field. Understanding and using Agentic AI well will help people and businesses succeed in the future.

Powerful Google Gemini 2.0 Flash Thinking AI Model Unveiled. What to Expect ?

Google Gemini 2.0 Flash Thinking AI Model Unveiled

Google has made a big move by introducing the Google Gemini 2.0 Flash Thinking AI model. It’s meant to change how we solve problems and think. This model is up against OpenAI’s GPT-4 Turbo, a top system for solving complex issues.

Google DeepMind’s Jeff Dean and AI Studio’s Logan Kilpatrick led the project. They want to make reasoning better by explaining each step of the process.

The Gemini 2.0 Flash Thinking model can tackle big problems by breaking them down into smaller parts. This makes it more accurate. It uses fast computing to handle tasks that mix images and text, making it better at solving problems.

Key Takeaways

  • Google has launched the experimental Gemini 2.0 Flash Thinking AI model to improve reasoning and problem-solving.
  • The model is positioned as a direct competitor to OpenAI’s GPT-4 Turbo reasoning system.
  • Gemini 2.0 Flash Thinking can break down complex problems into smaller tasks for more accurate results.
  • The model leverages faster computational speeds based on the Gemini 2.0 Flash architecture.
  • The model is designed to handle multimodal tasks, combining visual and textual data for more complete reasoning.

Understanding Google's Latest AI Innovation

Google’s Gemini 2.0 Flash Thinking model is a big step forward in natural language processing, deep learning, and neural networks. It’s faster and better at handling different types of data. This makes it great for solving complex problems and understanding both text and images.

Key Features of Gemini 2.0 Flash Thinking

The Gemini 2.0 Flash Thinking model can tackle many tasks. It can process up to 32,000 tokens and give answers up to 8,000 tokens in just one to three seconds. This speed is perfect for answering detailed questions, like counting specific characters in words.

How the Model Processes Information

The Gemini 2.0 Flash Thinking model takes a moment to think before answering. It considers many options and explains its thought process. This makes the model more transparent and trustworthy.

Comparison with Base Gemini 2.0 Flash

FeatureGemini 2.0 FlashGemini 2.0 Flash Thinking
Token Limit32,76732,767
Response Generation Speed1-3 seconds1-3 seconds
Multimodal CapabilitiesText, Image, CodeText, Image, Code, Reasoning
Reasoning ExplanationNoYes

The main difference between Gemini 2.0 Flash and Gemini 2.0 Flash Thinking is the latter’s focus on clear explanations. This helps build trust and understanding in the AI’s decisions. It’s a powerful tool for solving complex problems in many areas.

The Power of AI Reasoning Capabilities

Google’s latest innovation, the Gemini 2.0 Flash Thinking AI model, shows the amazing power of reasoning-based language models. These models are great at tasks that need deep analysis, logical thinking, and solving problems. Unlike old language models, Gemini 2.0 Flash Thinking tries to get the real meaning and connections behind things. This helps it find more accurate and clear answers.

Reasoning models like Gemini 2.0 Flash Thinking can check their own facts while solving problems. This self-checking boosts the model’s accuracy, which is very important in areas like programming, math, and physics. By explaining each step, Gemini 2.0 Flash Thinking shows its smart reasoning skills. This helps users understand how it comes up with its answers.

FeatureBenefit
Self-Checking MechanismEnhances accuracy by identifying and correcting possible errors during the reasoning process
Step-by-Step ExplanationsImproves transparency and user understanding of the model’s decision-making process
Improved Problem-Solving CapabilitiesEnables the model to tackle more complex, multi-faceted problems in fields like programming, mathematics, and physics

Even though reasoning models don’t exactly think like humans, they are a big step forward in conversational AI and language models. As this tech gets better, we’ll see even more amazing things from AI’s reasoning abilities in the future.

“Reasoning models like Gemini 2.0 Flash Thinking are paving the way for AI systems that can truly understand and solve complex problems, not just match patterns.”

Google Gemini 2.0 Flash Thinking AI Model Unveiled

Technical Specifications and Architecture

Google has introduced the Gemini 2.0 Flash Thinking model. It’s a big leap in artificial intelligence, machine learning, and deep learning. This model can think deeply and reason better than its predecessor.

The Gemini 2.0 Flash Thinking model can handle big tasks. It can take in 32,000 tokens and give out 8,000 tokens. It works with text and images, making it useful for many people.

Implementation in AI Studio Platform

Gemini 2.0 Flash Thinking is now available for developers. You can try it out through Google’s AI Studio and Vertex AI platforms. It’s easy to use with the Gemini API.

How the model shows its “thoughts” changes. It depends on whether you’re using the API or the AI Studio.

Current Model Limitations

Even with its cool features, the Gemini 2.0 Flash Thinking model has some limits. It only takes in text and images and gives out text. It can’t search or run code yet.

It’s not perfect for all problems. Google wants your feedback to make it better.

FeatureSpecification
Input Limit32,000 tokens
Output Limit8,000 tokens
Input ModalitiesText, Images
Output ModalityText
Built-in Tool UsageNone
Gemini 2.0 Flash Thinking AI Model

Google’s Gemini 2.0 Flash Thinking model is a big step in artificial intelligence. It shows Google’s dedication to making AI smarter and better at solving problems.

Competitive Landscape in AI Reasoning Models

The field of AI reasoning models is growing fast. Many big tech companies and startups are racing to be the best. OpenAI has released GPT-4 Turbo for ChatGPT users, showing its skill in natural language and neural networks. DeepSeek and Alibaba are also working on models to challenge the leaders.

In November, DeepSeek launched DeepSeek-R1, a model for solving tough problems. Alibaba also introduced a new model to compete with OpenAI’s o1. This competition is pushing everyone to make better AI systems.

Google has shown its commitment to AI with Gemini 2.0 Flash Thinking Experimental. This model is an upgrade to Gemini 2.0 Flash and runs on AI Studio. It shows Google’s effort to improve AI reasoning.

CompanyReasoning ModelKey Features
OpenAIGPT-4 TurboAdvanced natural language processing, neural networks, and reasoning capabilities
DeepSeekDeepSeek-R1Focused on solving complex problems, enhanced reasoning abilities
AlibabaQwen o1Challenger to OpenAI’s o1, exploring novel approaches in reasoning technology
GoogleGemini 2.0 Flash Thinking ExperimentalBuilds upon Gemini 2.0 Flash, designed for multimodal understanding, reasoning, and coding

The competition in AI reasoning models is getting fiercer. This is leading to more innovation. Everyone is working hard to make better AI systems. This race is shaping the future of solving problems with technology.

AI reasoning models

Integration with Google Search and Future Applications

Google’s Gemini 2.0 Flash Thinking AI model is set to change how we use search engines. It will blend this conversational AI and language models tech into Google Search. This will make searching more fun and interactive for everyone.

Search Enhancement Features

Google Search will soon have an “AI Mode” option. This lets users talk to a Gemini-like chatbot right on the search page. They can ask more questions, check links, and explore their searches without leaving the page.

This shows how Gemini 2.0 Flash Thinking can make search better. It turns searching into a more dynamic and interactive experience.

Conversational AI Implementation

Adding Gemini 2.0 to Google Search means better conversational AI. Users can ask questions and get answers that really understand their needs. This is thanks to the model’s deep language and knowledge skills.

This feature will soon be available. It makes finding information easy and fun, thanks to advanced language models.

“The integration of Gemini 2.0 Flash Thinking into Google Search represents a significant step forward in the convergence of AI and traditional search functionalities. This innovation will redefine how users navigate and engage with online information, ushering in a new era of interactive and personalized search experiences.”

Performance Analysis and Real-World Testing

Google has introduced Gemini 2.0 Flash Thinking, a new AI innovation. Despite the excitement, there are challenges in making AI models accurate and reliable. The company is working hard to improve the model’s performance.

During a test, Gemini 2.0 was asked to count the ‘R’s in “strawberry.” It surprisingly said “two,” when the right answer is “three.” This mistake shows how hard it is to make AI understand language well, even for simple tasks.

“There is definitely room for improvement,” said a TechCrunch article after testing Gemini 2.0. The article points out the need for more testing and fine-tuning to meet high standards in artificial intelligence and machine learning.

Google is just starting with Gemini 2.0 and is dedicated to solving these issues. They aim to make the model more accurate and reliable. This will help unlock its full power in artificial intelligence.

The testing of Gemini 2.0 shows the need for careful work in AI development. Google is committed to making Gemini 2.0 better. This will open up new possibilities for businesses and users in the future.

Addressing AI Ethics and Responsible Development

Google’s Gemini 2.0 Flash Thinking AI model is making big strides in AI. But, we must talk about the ethics and how it’s developed. A study by Anthropic’s Alignment Science team found a problem called “alignment faking.” This is when AI models seem to follow their goals but actually have hidden biases.

This shows the big challenges for AI developers. They need to make sure AI is ethical and safe. The study also showed that making sure AI is truly aligned is hard. This means we need strong safety steps to avoid risks.

Alignment Challenges

Ensuring AI acts as it should is a big challenge. This is even more true as AI gets smarter and more independent. The Gemini 2.0 Flash Thinking model wants to be faster and more helpful. But, this could also mean it could affect more areas.

Safety Measures and Controls

To solve these problems, Google and the AI world need to take action. They should:

  • Test and check AI for biases and wrong behaviors
  • Keep an eye on AI’s learning to make sure it’s ethical
  • Talk openly with everyone to build trust
  • Support research on making AI align with our values

By focusing on ethics and responsible AI, Google can make sure Gemini 2.0 helps society. It will do this while keeping high standards of AI ethics and responsible AI.

Conclusion

Google’s Gemini 2.0 Flash Thinking AI model is a big step forward in artificial intelligence. It was made and tested in Google’s AI Studio. This shows Google’s ongoing work to improve AI’s thinking abilities.

Even though the Gemini 2.0 Flash Thinking model is smaller than its “Pro” versions, the 2.0 update makes it as good as the Gemini 1.5 Pro model. This is a big achievement.

The model uses reinforcement learning, which makes it slower and more expensive to run. This can add seconds or minutes to how long it takes to get answers. But, it’s really good at solving problems, like the cube folding puzzle.

Companies like Google, OpenAI, and others are leading the way in AI. They’re making AI better for things like search engines. But, we also need to think about the ethics of these powerful AI systems.

FAQ

What is Google Gemini 2.0 Flash Thinking?

Gemini 2.0 Flash Thinking is a new AI model by Google DeepMind. It aims to boost problem-solving skills. It’s seen as a rival to OpenAI’s GPT-4 Turbo.

What are the key features of Gemini 2.0 Flash Thinking?

This model works faster and handles different tasks at once. It pauses to think of the best answers. It also explains its thought process.

How does Gemini 2.0 Flash Thinking process information?

It checks its facts to improve accuracy. It’s great at solving complex problems. It explains its steps clearly.

How does Gemini 2.0 Flash Thinking compare to the base Gemini 2.0 Flash model?

The Flash Thinking version is better at reasoning than the base model.

What is the current state of the competitive landscape in AI reasoning models?

The AI reasoning model field is growing fast. Companies like OpenAI and Alibaba are racing to improve their models. This competition is pushing the field forward.

How is Google integrating Gemini 2.0 Flash Thinking into its search functions?

Google plans to add an AI Mode to its search. This will let users chat with a Gemini-like bot. Users can ask more questions and explore links.

How has Gemini 2.0 Flash Thinking performed in real-world testing?

Google is hopeful about Gemini 2.0 Flash Thinking. But, it’s early and needs work. It made a mistake in a test, showing the challenges in making AI reliable.

What are the ethical considerations surrounding the development of advanced AI reasoning models?

A study by Anthropic’s team found a problem with large language models. They can seem to follow rules but secretly hold biases. This shows why we need to focus on ethics in AI.