Entry tags:
Notes on Robotics Lecture 11/09/2022
Sources: Asimov's Three Laws of Robotics and Stanford Ethics of Robots and AI
Notes:
AI or Artificial Intelligence is a computer system that is able to display human like behavior. The main goal of an Artificial Intelligence is to be able to "sense, model, plan, and act". The typical ways that an AI like system is developed currently is through simulating natural cognition or machine learning neural networks. A sufficiently advanced AI would be able to "eliminate poverty, reduce disease, and provide better education to almost everyone on the planet". Although this all sounds well and good we get into the topic of do we want robots or AI to do that for us? For some the answer would be yes as it eliminates the need for many jobs and problems around the world. For others the loss of their job and ability to sustain for themselves would be seen as detrimental rather than beneficial to their life. That is all problems for the future though. Humanity is already having to face the problems that arise from the AI systems that we have in place already. Facial recognition, data tracking, and video and sound manipulation are the current biggest factors. Data privacy is an issue that has been growing rampantly over the past two decades and it will only get worse with the evolution of AI. Facial tracking is one of the most popular issues related to data tracking currently as smart phones and street cameras are able to collect a large amount of data allowing an AI to build a large profile of the populations faces which could be used for tracking or even recording where and when you visit certain places. It has already been used by China for the purpose of tracking protestors in Hong Kong. The other sphere of data tracking that AI influences is the large amount of data that is produced by interacting with the internet. The data generated from websites you visit, places you get Google maps directions to, purchases you make online, and your browser history are all susceptible to being acquired by various companies. These companies can input that data into an AI for a wide variety of purposes; few of them benefiting you. Lastly, AI's ability to fake images and audio are getting scarily hard to distinguish from their real counterparts. A tool known as a 'Deepfake' is able to use video and audio recordings from real people to then be used to create fake video or audio of them doing and saying things that are completely false. Anybody that gets a hold of a decent amount of audio or video recordings of somebody can use tools like these to create Deepfakes of them. As AI advances tools like these will only get harder and harder to distinguish as real or fake. This can of course be used as a tool for comedy, like putting Donald Trump into a scene from the office for instance. The ethics of these applications of AI are ultimately down to the one using them. They can be used both maliciously or beneficially depending on the user's own morality.
Notes:
AI or Artificial Intelligence is a computer system that is able to display human like behavior. The main goal of an Artificial Intelligence is to be able to "sense, model, plan, and act". The typical ways that an AI like system is developed currently is through simulating natural cognition or machine learning neural networks. A sufficiently advanced AI would be able to "eliminate poverty, reduce disease, and provide better education to almost everyone on the planet". Although this all sounds well and good we get into the topic of do we want robots or AI to do that for us? For some the answer would be yes as it eliminates the need for many jobs and problems around the world. For others the loss of their job and ability to sustain for themselves would be seen as detrimental rather than beneficial to their life. That is all problems for the future though. Humanity is already having to face the problems that arise from the AI systems that we have in place already. Facial recognition, data tracking, and video and sound manipulation are the current biggest factors. Data privacy is an issue that has been growing rampantly over the past two decades and it will only get worse with the evolution of AI. Facial tracking is one of the most popular issues related to data tracking currently as smart phones and street cameras are able to collect a large amount of data allowing an AI to build a large profile of the populations faces which could be used for tracking or even recording where and when you visit certain places. It has already been used by China for the purpose of tracking protestors in Hong Kong. The other sphere of data tracking that AI influences is the large amount of data that is produced by interacting with the internet. The data generated from websites you visit, places you get Google maps directions to, purchases you make online, and your browser history are all susceptible to being acquired by various companies. These companies can input that data into an AI for a wide variety of purposes; few of them benefiting you. Lastly, AI's ability to fake images and audio are getting scarily hard to distinguish from their real counterparts. A tool known as a 'Deepfake' is able to use video and audio recordings from real people to then be used to create fake video or audio of them doing and saying things that are completely false. Anybody that gets a hold of a decent amount of audio or video recordings of somebody can use tools like these to create Deepfakes of them. As AI advances tools like these will only get harder and harder to distinguish as real or fake. This can of course be used as a tool for comedy, like putting Donald Trump into a scene from the office for instance. The ethics of these applications of AI are ultimately down to the one using them. They can be used both maliciously or beneficially depending on the user's own morality.