Generative AI (GenAI) - What Is It & How Does It Work?

Generative AI (GenAI) – What Is It & How Does It Work?

Generative AI is a new and trending method of creating Content that operates using very large datasets. Come learn more. Generative AI is relatively New Technology. And, this GenAI technology is going to be around for a long long time. The ceilings this technological innovation has broken has been likened to the technological and digital Revolutions such as when Electricity. Information and Communications Technology (ICT), and the Internet were invented. So before we look at what GenAI is, lets see why there is such a Rave about this new technology.

Why The Rave about Generative AI (Gen AI)?

Many people are intrigued by what Gen AI is able to do. A software program is able to converse with me – Alexa or Siri. My mother has Alexa, and I sometimes engage Alexa in conversations. It amuses my each time. On the other hand, I marvel at the genius and intuitiveness of programmers, who come up with the codes that birthed Gen AI. Especially because Gen AI has eased up our lives in so many ways. Unimaginable what the mind can conceive. But, not only has Gen AI eased up our personal lives. It has also boosted our businesses and marketplace in general. Medicine is not an exception. More will be discovered in this realm. Education and Mental Healthcare are no exception, either. Nor are Marketing and blogging systems. Everywhere and anywhere we have the impact of Gen AI. This is definitely worth raving about. Don’t you think? Sometimes I wonder how we can use Gen AI to support people to live healthier lifestyle. That I have to sit on for a while. Maybe it come to me in now time. Once that happens. I shall be chronicling about it. Shalom (peace)!

For a long time, I have been curious about AI, ML and Gen AI. Let’s explore together… There is definitely a hierarchical relationship between Artificial Intelligence (AI), Machine Learning (ML), and GenAI.

Here is a comparison between AI, ML and GenAI…

AI: Performs tasks that typically require human Intelligence (HI). AI mimics HI across applications

ML: Makes decisions and prediction based on provided datasets. ML also makes prediction and decisions on new previously not seen data.

GenAI: Generates new data based on provided information. GenAI is able to produce new data that is Not part of the Original dataset, but have similar characteristics.

Generative AI has the capacity to be creative. Examples of some popular GenAI are ChatGPT (for text generation), Midjourney (for images), and DALL-E (for generated images).

So, Before We Look at What Generative AI really Means, Let’s Watch this Video

What is Generative AI?

Generative AI mimics human creativity, which is markedly different from discriminative AI that differentiates between types of information. Generative AI has the capacity to repurpose, and rephrase different types of information, Some examples GenAI repurposing include,

Answers to inquires
Videos
Music
Webinars
Texts
Computer codes
Images and so forth

How does Generative AI Work?

Generative AI algorithms are built with Deep Neural Networks, which is a branch of Machine Learning.

First,

A large amount of data is fed into a Large Language Model (LLM) which has large amounts of pretrained data. Pretrained information are usually from different sources such as data from companies ledgers, books, or web pages, and so forth. Therefore, the type of data that needs to be generated determines what information is input into the LLMs.

Next,

The data fed into the LLM has to be made understandable to the software. Sentences, and other types of data sequences are transformed into Vector Embeddings. Large Language Models use Transformers to convert these sentences and data sequences into Numerical Representations known as Vector Embeddings. It is within the vector embeddings these sentences, and words are converted into numbers in such a way the integrity of relationships and meanings of the information inputted is maintained.

Third,

The data sequences in the Vector Embeddings are converted into Vectors. These vectors are then organized according to how similar they are to each other. This grouping helps to determine how words are related. This step is very important because the effectiveness of this classification of words will determine how close the output produced is SIMILAR to the input (i.e., training) data.

Fourth,

The main idea is to get to a point when a Model can output intelligible and natural sounding and looking data. To accomplish such results, the training data must go through many computing processing stages. The model’s learning process is largely automated. However, people will have to continue to fine tune the training information to ensure it is accurate. This means the usefulness of the Gen AI output depends on quality and comprehensiveness of the training data. Also, the processes used to train the Model, the Model’s architecture which is neither too simple nor complex are fine tuned, and so are the instructions users give to the Model to produce output. These instructions are called Prompts.

Finally,

When the training data has been fine tuned, and processed many times, the model’s output Sounds and Looks Natural. This means the output then looks like a person wrote the text. And sounds as if a person is speaking. An example will be talking or texting with a virtual assistant. This is ingenious, isn’t it?

Creative Uses of Generative AI

Make ecommerce business more efficient
Speed up the discovery of new medications
Enhance creativity and content development, including art and images
Increase employee productivity
Inspire new ideas and innovation across different fields
Composition of music
Can write about everything and anything
Enhance online customer experiences, including improving customer support
Make process optimization faster

Potential Downsides of Generative AI

Original Image generation may be unrefined, which can be problematic
You cannot trust what GenAI tells you a 100%. You can trust GenAI, but you may have to verify
Unintentional plagiarizing of an author’s work can easily occur
Increase the potential for data breaches at the workplace
May decrease the financial value of human created content
Malicious people could abuse this technology and create all sorts of disruption
Could encourage students to rely on GenAI generated essays, rather than develop their own writing skills

My Thoughts

It is obvious that Generative AI has come to stay with us. Especially because it is opening up mind blowing and limitless possibilities. But, will it take over from humans at the workplace, educational institutions, marketing, or medicine, mental healthcare, and healthcare? I am not sure that will Ever happen. At least, not in the near future.

Best 4 Everything Content

Be sure to press [CTRL] & [D] to bookmark us!