The art of building new content: Generative AI By Aditya Abeysinghe

The art of building new content: Generative AI

By Aditya Abeysinghe

Is tracked data changing consumer behavior? By Aditya Abeysinghe

At present, regenerating images and audio from existing images and audio is a common activity. Most images are edited to attract viewers and audio files are edited to enhance listening quality. However, creating new content from existing files, especially generating images and other media files has been a trend that sprung recently. This process of using artificial intelligence (AI) to generate new content is known as generative AI.

What is the use of generative AI?

One of the uses of generative AI is to generate synthetic content inexpensively. For example, think about photos of models for image recognition. Unless these photos are publicly available, they can rarely be used for commercial purposes due to copyright issues. However, with generative AI, new images can be generated with machine learning for humans that didn’t even exist. Thus, generative AI can generate content without privacy issues with minimal effort and expense.

Another use of generative AI is the usage of robots as helpers in tasks such as generation of sample content. For example, several software are at present being tested on the consumer space for generating texts for articles, essays and short stories. These help people build stories with help from robots and then change what is deviating from their original needs. Many plugins are now available to automatically correct documents on tasks such as proofreading and editing. These plugins use texts from a collection and then display words or sentences to suit the text available and either correct errors or enhance the wording.

Generative AI is also now tested in various other areas such as discovery of drugs. Neural networks are used to discover proteins and then simulate their actions with various molecules to discover drugs. These drugs are then tested on animals and released for human use.

What are the limitations?

The main limitation with generative AI is that they are hard to control from a machine learning viewpoint. The main reason for this is that these models are still in their infancy stages and are fine tuned for better results. Many models fitting the purpose of a solution can be developed. However, evaluating these models in terms of the desired result is hard. This is especially dangerous in areas such as drugs as explained because the end result is being tested on humans and animals. Therefore, generative AI should be cautiously used in such areas.

Another area that has been criticized is the use of generative AI for illegal activities such as marketing, where fake images and videos of popular actors and actresses are generated as illusions for viewers. Many illegal media have been generated and commercialized for generating profits based on such activities and many such cases have been banned from releasing to public.

Image Courtesy:

Comments are closed.