|
|
Can you see where can it cause problems? What are the consequences of not checking whether the machine learning program meets moral standards? William Tracy: First of all, pay attention to the art world. They are those who have less money and have the ability to change policies. The company will file a lawsuit through a lawyer and protect themselves from harm. What will a lonely artist do, his photos are re -created and sold by artificial intelligence, or a writer has three latest books cottage versions? First of all, people who create machine learning procedures should certainly bear part of their responsibilities, but are they using it to re -create products that bring income to others?
Does the person using the product guilty, because the product is provided for free, and there is no rule that what they do is immoral or illegal? Morality should be the first stage of any new technology design. Think abo Argentina WhatsApp Number ut how to use it, where to use it, and who it can affect. How does user privacy machine learning model bring risks to user privacy? William Tracy: The machine learning model is built on a large number of pre -existence data. This data can be real or generated. When using real data, accurate data may enter the output of the model. Keep in mind that there is no "intelligence" behind the content of each layer of the model.
All this is based on mathematics, so there is nothing to prevent the transfer of private information, unless humans add expedient measures in the model. An example in real life is also a way for people to find the image generation model from the Internet. USE. data privacy when training models? William Tracy: The best way to avoid data privacy issues is to do not use private data when training models. Then it cannot pass through the model layer and come out of the other side.
|
|
發表於 2024-3-5 12:19:48