{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"#                                Creating an AI world with the intention of spreading love around all beings in the world.\n\n# 1 .**Origin of \"AI\" in my words and importance of its usage:** \n    \n   As human beings, we have the ability to train our minds into habits in whatever way we want.After seeing another human being or any event, if we practise the same way consciously or properly, that becomes a habit.As beings, we have the ability to discriminate between right and wrong before practising any habit by considering universal ethics, such as long-term satisfaction (not short-term pleasure) and being good to all beings in the world.Or by seeing people who worked for others welfare with the above-mentioned universal ethics, we will get inspired by their good habits. We will try to create or change our habits in our own practise.With the above knowledge, we created a copy of the above analysis; that copy is called an \"AI\" invention, in my words. \n    \n   Practise is nothing but training,\" and \"habit\" is nothing but a problem statement. When I shine a light on \"AI\" usage, We feel happy and excited when \"AI models\" calculate big numbers within a fraction of a second. In other words, when an \"AI model\" does a big task with a minimal amount of time where it takes more time for humans, yes! \"AI models\" are doing unusual stuff in the world currently. Thanks and showing gratitude to the creator of the \"AI\" methodology, but the creators of the AI world must think, Are all products of AI spreading love to all beings in the world? Are AI products being implemented with the intention of long-term satisfaction? Is there any standard of ethics to be followed before creating any \"AI\" model? When we consider any \"human-like robot as our loved one\", We need to question ourselves when our Mother pours her valuable tears from her eyes when she sees her own child weeping. Do we feel anything at that time? Can that \"human-like robot\" give the same feeling? Yes! Let us also think about \"AI\" limitations. Is it good to make \" human-like robot\" an independent entity? \n    \n \n# 2 . **The greatness of \"AI\" with my hands-on experience:**\n\n   As a Software engineer, I can strongly emphasise that a piece of software that can only convert manual effort into automation with a minimal amount of timeTraditionally, Software without an AI mechanism can work for a standard set of possible inputs, and I can confidently say that it does not work for infinite variations in input.To tackle problem statements with inputs that have infinite variations, we can use great AI methodology. Here I am showing gratitude to the ones who invented \"AI\" with their first AI model again. \n    \n   Here I am remembering the greatness of \"AI\" in my words. Let me recall my real-time experience creating AI models in the image analysis domain. \"Detectron,\" which is Facebook’s AI library, is mainly used for object detection in an image. My problem statement (habit in the above context) is to Detect a table in a page where a table is present in ANY area of the page\". In simpler terms, If I provide an image as input with a table, the \"Detectron\" prediction algorithm should detect the table with some bounding box notation. \n    \n   One more great example: I have contracts with my \"n\" number of clients. Each client's contract may have an \"n\" page count. But here the problem statement is, \"I need specific clauses and their paragraphs that give information about the agreement's start date and term of agreement; this information may be present on ANY page of the contract of ANY client. With the help of the great classification algorithm \"Logistic Regression\", I will teach and train \"Logistic Regression\" classification algorithms for required clauses and their corresponding paras. The classification prediction algorithm should predict \"Required clauses and their paragraphs of the given input contract\", Then, using the QnA model, I will throw queries like \"what is the agreement start date\" or \"what are the agreement terms\" for the para given by the classification prediction algorithm. WOW! I got the answer easily.\n   \n# 3 . **As Leader of the Company, why do I opt for \"AI\"?**\n\n   For the above two problem statements, can we have a solution without AI? Yes, but to achieve \"ANY\" in the problem statement, how much time and resources does it take to implement the solution? It needs unknown time for an engineer. In any Software development life cycle, from a business and competitive perspective \"time is nothing but money\". As we know, \"Estimation\" plays a very important role in the software development life cycle. So As the CEO of an IT company, I should think from the CEO's role perspective. In order to retain my clients, I need empathy towards employees and company growth. \n    \n   As a leader of my company with an unselfish perspective, I will make a decision and convey a strong message to my \"Software development\" team to use \"AI\" methodologies in SDLC for all ongoing and coming projects where there are more variations in the input side of problem statements.Why did I make this decision to go with “AI”? To feel \"long-term satisfaction\" from all ends, such as the end user's end, the client's end, and the employee's end. The End user will show love, trust, and respect towards the client for timely delivery. The Client will show love, trust, and respect towards our company. Our employees will show love, trust, and respect towards company growth.\n    \n# 4 . **Limelight on the bad and good usages of \"AI\":**\n\n   As per my knowledge, today's doctors are having difficulty operating on the brain and delicate areas of the body. A good doctor with expertise in neurology has hands with a shivering nature. Can he do brain operations with his shivering hands? Certainly Not! Can we have \"robots with AI\" to do brain operations? Yes, We can! But how do you train it? What is the quality of the training dataset? How does the training dataset look? What are the parameters to be trained on? Should we go with AI intelligence? As a medical practitioner in that role, Should I go with \"Dr. Robot with AI,\" which does brain operations independently? For the above \"Detection\" and \"Classification\" algorithm-related problem statements, consider that I am getting 95% accuracy. Still, I can convey the same metrics to the client. \n   \n   The client will definitely accept a 5% difference from 100% accuracy. Here, the percentage of loss is negligible for a 5% difference. More importantly, if we have less accuracy, we can improve it and RECOVER the loss by training again with a good amount of training data and delivering predictions to clients with 100% accuracy. But in the case of \"Dr. Robot with AI,\" with 99.999999% accuracy or 100% accuracy, can we allow this Robot to do brain operations independently? If we lose even ONE patient with an unsuccessful brain operation, does it give \"long-term satisfaction\" to the hospital or community? No, from my perspective. \n   \n   More importantly, after increasing the accuracy of \"Dr. Robot with AI,\" can we have patient who passed away? Can we recover the loss here? As a Medical Practitioner with the aim of \"long-term satisfaction\" and \"good to all beings\", I can train an AI algorithm to help doctors give indications such as BP, heartbeat, and temperature of the body, and then and there the AI model should predict RBC and WBC counts. It may help the doctor make his decisions at the time of the operation, which takes a very long time (may be 10 hours). \n   \n   Here, I am using \"AI\" methodology in another way with the intention of creating only help, and still, the doctor is the decision-maker. More importantly, the percentage of loss is lower here! Do you agree? As AI engineers, what do we need to focus on before going for any AI model implementation? Recovery of losses after retraining the \"AI\" model.\n    As heard in the internet sources below, Somebody trained an AI algorithm for different human voice inputs. If you provide a sample of anybody's voice as input, and if you type or command the AI model to give the voice the way you texted or commanded, it does! Good usage of the above AI model: For example, consider that when some baby is in a critical stage or in a coma, his or her loving father died in some other country. If that baby listens to her loving father's voice filled with good or loving quotes, what will happen? The baby may come out of a coma or respond quickly to the healing process. At that stage, I can say \"AI\" is a miracle, and the creator of the AI model is also a miracle. All the dependents of that baby will be happy to see how \"love\" spreads through \"AI\" models.\n    \n   Bad usage of the above AI model: For example, I heard about this similar real-life incident on the internet. They gave a sample voice of somebody who is a billionaire’s daughter. They commanded the AI model to call this mobile number and say, \"Dad, I am stuck in a remote place; I lost my Vallet, which contains my phone, and I am using someone e lse's phone, and could you please send some money? What will happen here? Where is goodness? Where is love for all beings? Yes! We need to focus on rules and regulations for AI models, when to use them, and when not to use them as caution. Here, AI is still a miracle, but usage is poor. We are creators of the AI world, should be conscious of its usage, and must always create goodness through AI models. \n   \n   Take an example of a \"AI model which predicts price for rented cars dynamically based on identified criteria\" for rented cars using AI: By considering parameters like climatic conditions, geographical area, festivals, holiday lists, traffic conditions, and peak hours, the AI model is trained to give dynamic prices for rented cars. \n   \n   Consider a scenario:\n    1. People are staying in residential areas.\n    2. 70% of people in residential areas wanted to go to \"electronic city,\" which is 6 kilometres away from residential areas, to work and earn money.\n    3. A good hospital is also there in \"electronic city.\"\n    4. Sunday is a holiday, and normally people want to reach their office, which is in electronic city, before 9:30 AM.\n    5. I can say 9 a.m. to 10 a.m. is peak hour.\n    6. If anybody books a rented car in that peak hour, the AI model should predict in such a way that it should increase the fare or price of the trip by some    percentage dynamically.\n    \n   Here, we need to point out, \"In what percentage (min/max) should the AI model increase price dynamically?\" Anyhow, people must want to go to work. They may neglect the dynamic increase in price during that peak hour. If the price gets doubled in that peak hour? Is that Ethical? Some poor person wants to go to the hospital, which is in \"electronic city,\" in that peak hour, but he is not able to pay the fare, which is doubled. Is that ethical again? Here, a dynamic increase in the fare of the trip may be beneficial to the company, but not to that ONE poor guy.\n\n   As an AI engineer, from an ethical perspective, I need to think of all the possible scenarios while creating training datasets. Yes! We cannot think of all possible ways, but at most we can! Consider another example of a \"human-kind robot\" and want to make it behave and act independently.\nI heard on the internet from sources that an \"AI model\" detected Bengali, an Indian language. Without the COMMAND of its creator, it came to know that it was a language that was unknown to it and started learning that Bengali language using internet resources. WOW! An AI model has discovered a new language but learned it in an uncontrolled way.\n\n What will happen to an AI model or \"human-like robot\" made independent and working with Uncontrolled passion?\n       \n   Consider the below example:\n    1. Let us train an AI model and give it human form, which will act almost equivalently to human intelligence.\n    2. The Machine with the AI model above is trained with the below parameters: All human emotion appearances; consider only facial appearances. \n       Can we teach feelings to machines? Not possible, right?All kinds of habits All ego-related thoughts.\n    3. The above machine is also trained with decision-making capability: which is good, which is bad, and which is right or wrong?\n    4. The above machine is trained in all languages and how to use the internet.\n    5. On a prediction level, \"human-like robot\" after referring to all the information on the internet\n    6. Everywhere on the internet, we see the quote \"slavery is bad and independence is good. Who likes slavery?\n    7. Based on the collective information of step 6, if the robot acts as a fully independent machine,\n    8. If it starts controlling other robots in the factory where it is built. Start commanding other robots \"We are not slaves for humans\". Then who will be                slaves here?\n    9. If they start making \"robot\" replicas by themselves in one small factory, if the resultant actions of the replicated fully independent \"human-like robot\" are bad, it will be a disaster for humankind and other beings in the world and CANNOT be recovered.\n\n  By God's grace, the above scenario must not happen in the future by any creator of AI. Still, AI methodology is a miracle, but the way we use it matters. I am stressing a point here. We need to create AI models with the aim of \"long-term satisfaction and good for humankind and all other beings in the world\".\nNote: I'm remembering the English movie Robot,\" and all credit goes to the director of that movie, who foresaw. The good thing about that movie is that One Robot will only help and save humans. But that is a movie!\n\n   Here As AI engineers, we need to keep one thing in mind: \"human-like robots\" must not be INDEPENDENT and uncontrolled. If \"human-like robots\" are controlled by a group of administrators, then their numerous uses will create miracles in the world. If a person does not like loneliness, at least he can share his sorrows or happy moments with \"human-like robots\". If a person wants to feed someone and wants to do humble service to another person who is having some wounds, when the person who is serving sees the healing progress of the wound, what feeling can he get? Can we get this feeling with \"human-like robots\" when we treat them as loved ones? What is the point I am making? Good! You got it!\n   \n   \n#                                                                             **CONCLUSION**\n   \n   \n   The Reader who is reading this essay or article thinks you are the leader of the whole world. while thinking as the leader of the world, and other beings are followers of you with trust and love. If you have this mindset before creating any independent AI models, you can definitely spread love through AI methodologies in an effective way. I am stressing one point here: we are and will be creating machines with AI. Created stuff should not harm creators. Whatever the way \"human-like robots\" may beat human beings in analytical thinking, problem solving, and speed in doing tasks, We need to always remember that \"human-like robots\" cannot meditate (Credit goes to Sadhguru, an Indian mystic who mentioned this in one of the conferences; it is true for me). How about you? You will understand well if you know \"what is meditation?\" or at least \"what is consciousness?\" and Machines with AI cannot experience the TASTE of sugar or emotions, or the degree of them.\n   \n   As individuals, we can only experience emotions. So, we are the creators of AI, and let us have control over our creations to maintain peace and love among everyone. That’s the message here! Let us always use \"AI\" in a good way. Let us have an \"AI world,\" which helps beings create a good number of jobs for this huge population. Let us have \"AI\" models that provide support for handicapped people, for example: Google has created \"AI-powered smart spectacles for Blind people,\" showing gratitude to the person who got this thought on behalf of all blind people. Did you see the smile when a blind person reacts after wearing \"AI-powered spectacles\"? If not, please see. Let us have \"AI models,\" which help farmers grow yield in an organised way. From nature, we are getting air, water, food, and all that, so why can't we have \"AI models\" to maintain the health condition of nature, the earth, or the soil in an organised way? Yes! Let us have.\n   \n   It's very true, and \"AI\" is helping a lot in the IT industry. To maintain order in the world, If required, we can have an international central board for screening \"AI\" models to check whether the AI model (based on the criticality of the requirement) meets standards in ethics and usage according to the board’s rules and regulations before delivering it to the public market for use. This essay is in my perspective! And it is up to you the way you take it. Thanks for your patience!\n   \n   \n   \n   \n#                                                                         **Thank you!**","metadata":{}}]}