AI is now hyped more than ever and with this trend is better to look at it to get a better understanding. Developers are now rushing to put a lot of efforts in it such as marketing to promote it , it services and the enormous work it can do.
{getToc} $title={What This Article Entails } $count={false}
Artificial Intelligence (AI) is the simulation of human like intelligence by computer systems. Ai requires special hardware and software to training the computer to mimic human intelligence. No specific programming language is used but Java, JavaScript, Python, C+ and julia are popular when it comes to designing Artificial Intelligence.
Types of Ai
Artificial Intelligence can be grouped into various types: Based on capabilities and based on functionality. Let's take a look at them.
Based On Functionality
1.Purely Reactive Machines
Purely reactive machines are the most simplest and basic type of Artificial Intelligence. These do not have memory or rather it does not store past experience for future actions. they only focus on current problems and try to react to them with best solution. An example is the IBM's Deep blue that was commonly used in the 1990's. it was used in the chess program to beat the Russian chess grandmaster Garry. Another example is the Google's AlphaGo.
2.Limited Memory
Limited Memory Machines can store past experiences or data for a limited time only. These can only use this data for a limited time or period. A great example is self driving cars. They store only limited memories such others cars speed and signs so it can use it for navigation.
3. Theory of Mind
The theory of Mind should be able to understand human feelings, emotions and interact socially like humans. This AI does not exist yet but there are a lot of efforts to improve and create one such like.
{inAds}
4. Self Awareness
In this type AI systems have a sense of self consciousness which can understand it's own current state. These will be smarter than the human brain. Self Awareness AI does not exist yet and it is just an hypothetical idea.
Based on Capabilities
1.Narrow or Weak AI
Narrow AI is a AI that work on a specific task. It is dedicated to it's task and limited to it. It cannot perform beyond it's limitation or fields since it is dedicated to a specific task. Therefore it is termed as weak AI. It is the type of AI that is most currently and available to most users. Weak AI can fail Unpredictably if it goes beyond its limits. An example is Apple's Siri which operates with a limited pre-defined range of functions. Other example is playing chess, personal recommendations and self driving cars.
2.General AI
General AI often referred to Artificial General Intelligence (AGI) would be able to perform any task a human could do. This type of AI would be smarter and be able to think like a human. To do so, AGI would need the ability to reason across a wide range of scenarios to understand complex problems it was not specifically programmed to solve. This type of AI does not exist yet but a lot of researchers are focusing on building machines with General AI
3.Super AI
Super AI is intended to be intelligence of systems that would surpass human intelligence and can perform any task better than human with cognitive properties. This would be an outcome of general Ai. This type of AI does not exist yet and it is just an hypothetical idea.
How Does Ai Work?
AI works by collecting large inputs of data which is used for training. First a lot of input is collected in form of text, audio, video and more. The collected data is then categorized into two; those that can be read by algorithms and those can not. The AI is then trained to decrypt data or understand it using the the patterns it has been programmed to learn until it develops similar patterns. From there one the AI is let to decide whether the data input is fed matches the previous one or not. If it fails to do so, the training may be repeated under different circumstances or environments or may be reprogrammed or adjusted it's algorithms. If the AI passes it can proceed to the next step of completion and released to the market. AI can therefore use input data to predict the next outcomes. For example an AI model can be trained with thousands of images labelled "dogs", "Cats" or "cows" so that it can recognize it immediately and it doesn't confuse it with other animals. Ai can also be fed with various human texts such that it can understand it and be able to respond to it. These can result in make of models such as ChatGPT for natural language processing (NLP) or Stable Diffusion for image generation. The efficiency of AI would depend on the variety of data and the amount of training data. For example an AI trained with less variety images of "birds" may have trouble at identifying birds. This would require to have machines with great graphics card for high resolution images. Ai is trained to identify objects in different resolutions that is higher and lower quality inputs.
Advantages and disadvantages of Ai
I may not be able to cover the full extent of the advantages and disadvantages of AI since there quit a lot. But here are some of them.
{inAds}
Advantages
- Consistency: AI application have delivered consistent and reliable outcomes. Analytic's tools use AI to process extensive amounts of data in a uniform way while retaining the ability to adapt to new information through continuous learning.
- Time Saving: AI have been used to automate operations thus saving a lot of time thus giving more room for creativity and increasing creativity. This is helpful where routine data entry is done.
- Excellence in detail oriented tasks: AI is great at performing tasks that may involve identifying similar patterns and relationships in data that may be overlooked by humans.
- Safety: AI and robotics have been used to perform hazardous tasks that might be harmful to human beings therefore reducing the risks and creating more time for creativity and productivity gains.
- Customization and personalization: AI can be used to enhance personal interactions and content delivery on digital platforms. An example is recommendations engine on platforms like Netflix that interacts with your preference and are able to suggest the next movie or TV show for you.
- 24/7 Availability: AI can be accessed at any time as they do not need time to rest and does not get tired. This can be efficient in Chatbot AI that can talk and assist customers at any time of the day or night.
- Scalability : AI systems can handle amounts of Data and inputs that are growing. This can be useful in business that could grow exponentially with time.
Disadvantages
- High costs: The cost of developing an AI is very expensive. Building an AI model requires a substantial investment in computational resources, infrastructure and the software to train the AI model and to store its training data. The costs does not end here there is still those incurred in retraining the model and marketing.
- Shortage of technical know how : There is a significant shortage of professionals trained in AI and machine learning. Also there has been an increase in the development of AI but less professionals are there.
- Technical Complexity: Developing operating and troubleshooting AI systems requires a lot of technical knowledge know how. This knowledge is quite different from the developing non AI systems and applications.
- Difference in generalization: AI performs well in doing specific tasks which has been trained for but when it comes to general tasks it becomes a complex task for them to achieve. This is because AI are required to change their algorithms and adapt to the changes to solve a particular problem that it was initially programmed to solve.
- Job Displacement and loss of Jobs: Lately, there has been a lot of job displacement and loss of jobs. AI has been used to automated workflows systems in industry. Also human are been replaced by machines which can perform regular tasks without getting tired and saving a lot of resources.
- Security Vulnerabilities: AI is exposed to a lot of security implications such as cyberthreats, data poisoning and adversarial machine learning. Hackers can extract sensitive information from the training data of an AI model. People may also trick AI models into providing incorrect and harmful output. This can be concerning in sensitive sectors such as financial services and government.
- Legal issues: Using AI to analyze and make decisions based on personal data has serious privacy implications.
Application of AI
AI in Entertainment
One of the popular AI that almost everyone has interact with is the Recommendations engine. AI can look what you looked at before, liked and how you have interacted with it and can suggest similar content such as movies, music, artists and products.
Second is the creative AI. This is more of an assistant AI that help digital creators. For example it may help artists to generate beats and correct certain features while they focus on their goals and visions. Another example is the the graphic designers, AI can help in removing backgrounds, suggest colors and fonts that may attract customers or a targeted audience.
AI in Health Care
AI can be used to examine pictures of MRI's and x-rays by using special algorithms to identify tumors and broken bones more accurately. Doctors can therefore focus on finding the right treatment for the patient.
AI can also be your personal health coach. AI can look at your previous medical records and activities to see how you reacted to certain treatments and can therefore come up with a specialized treatment or schedule for you. The treatment then becomes more efficient. AI can remind of times when to take medications, time for exercises, time to sleep and time for health breaks.
AI in gaming
AI are used to create NPC's (Non playable characters) in games. The AI uses stimulation and clever algorithms to create NPC's that look real and act like people in real life. It makes the gaming feel realistic and more immersive.
AI can be used to create maps, levels and backgrounds without the help of human beings. This can make game become more fun as your proceed to other levels, you meet other characters, worlds and different challenges.
AI in finance and Banking
AI can be used to keep an eye on the banking transactions all the time. They act as detectives who watches every transaction and can be used to prevent fraud. They can detect something fishy like depositing or withdrawing large sums of money which can help raise alarm and help to stop fraud.
AI can be used to automate trading.AI can analyze swiftly various markets trends and can be used to buy and sell stocks. This boosts trading strategies, making investments more efficient and profitable.
AI in Social Media
AI can act as social media guide. It watches what you do on social media like videos you watch, liked posts and ads. It then suggest similar content and ads that you might like. It act as someone who knows and understands you taste and can recommend the same making your experience on social media more enjoyable and personalized.
AI chatbots and virtual assistants act as digital helpers. They can answer your questions like human at any time 24/7. So you do need to wait hours to get a human reply. For complex situations it can forward your issues to support team for help and give your appropriate feedback.
AI in Travel and transport
AI plays a crucial role in optimizing travel routes, be it for parcels deliveries, public transportation, or personal trips. It efficiently calculates the swiftest and most economical paths from one point to another point, resulting in reduced travel time, minimized fuel consumption, and cost savings. Essentially, it serves as a pocket-sized travel advisor, enhancing the speed and budget-friendliness of your journeys.
AI can also be used to scan for bags and people in airport. It uses special kills to scan people quickly and ensure safety in the airport for everyone.
AI can also detect when cars or vehicles are about to have problems before they happen. This helps in fixing issues before happening. It can also be used to troubleshoot problems and suggest fix for vehicles.
{inAds}
AI in Education
AI can be used as teaching assistant for teachers. It can help them create quizzes , study materials and lesson plans. This makes education more easier as teachers are left with more time to interact with students and understand them better.
AI can also be used as virtual assistants for students. They can help students learn better by providing visual and audio presentations. This becomes even more efficient as students can access them any time they are stuck. It also relieves pressure on teachers as they can answer common question and provide study materials leaving teachers with time for personalized teaching.
Ethical issues regarding Ai
One of the most ethical concerns from unemployment. AI has been used to automate workflows and a lot of people have lost their jobs. Second is self driving vehicles, Drivers are losing their jobs to this. It is almost inevitable that AI will take your job. The biggest concern is how will the wealth gained from AI workflow be divided? Will the wealth be divided to those who lost their jobs or will go to the single individuals who use AI to do most of the tasks?
Second concern is about data integrity, AI can be used to create images and videos that look very much realistic. This can be used to create deepfakes. How do we differentiate between actual events and faked events? Understanding that deepfakes can be used to access someone's private data such as bank information and other private data. In 2015, a bot named Eugene Goostman won the Turing challenge for the first time. In the challenge, human raters used text input to chat with an unknown entity, then guessed whether they had been chatting with a human or a machine. Eugene fooled more than the half of the human raters into thinking they had been talking to a human being
AI reinforce what it has already learned. meaning they are highly dependent on the training data the have been given. The training data is selected by human beings and the potential for bias is almost inherent and therefore must be monitored closely. The question whether will be better or worse lies in its training data.
Governance and Regulations of Ai
The risks posed by AI have the potential to create social, economic and ethical implications. Ignoring the risks could result in serious implications and hinder technological process as well other processes. This is where international laws play a crucial role in setting goals and standards to reduce risks depending on their level and potential damage.
The European Union (EU) has addressed AI governance. The EU's General Data Protection Regulation (GDPR) has imposed strict limits on how companies can use consumer data, affecting the training and functionality of many consumer facing AI applications. The EU has approved the AI act which seeks to provide a legal framework for the development and governance of AI. The AI act imposes varying level of regulations based on their riskiness with areas such as biometric.
In USA, the White House of the Science and Technology Policy published a " Blueprint for an AI Bill of Rights" in October 2022, providing guidance on how to implement ethical AI systems. In October 2023, President Biden issued an executive order on the topic of secure and responsible AI development. The order directed federal agencies to take certain actions to assess and manage AI risk and developers of powerful AI systems to report safety test results.
Final thoughts on Ai
AI can be good especially when used to simplify tasks and create more time for creativity. Despite this AI can be used to perform harm which can pause ethical issues. Regulation and Governance is therefore necessary. Regulation of AI would not be easy since AI are developed for different various purpose and the rapid evolving technology makes it difficult but still the need for new laws to regulate the different AI should be formulated and implemented.
