What is Generative AI, and How Does it Work?

5 min read

Here’s what you need to know about tools like ChatGPT and Midjourney.

Generative AI has taken the world by storm. Tools like ChatGPT, Google Bard, and Midjourney are transforming whole industries, doing much of the heavy lifting that knowledge industry professionals previously did by hand. Google has even rolled out an AI writing assistant called Duet AI for use in Gmail and Docs through its Google Workspace Labs offering, making it easy for users to auto-generate text in emails and documents. As artificial intelligence advances, expect even more productivity tools to roll out generative AI tools to help knowledge industry workers achieve more output in less time. Here’s what you should know about how generative AI is changing white-collar work.

How Generative AI Works

Generative AI is a kind of artificial intelligence that can create – or generate – new content by leveraging existing content. Generative AI can create new documents, images, or even videos, using pre-existing assets as its inspiration.

Generative AI leverages Large Language Models, a type of neural network, to learn from previous examples of various types of content in order to generate its own new content.

The applications of generative AI are broad, and are growing every day. Generative AI can accomplish anything from drafting emails and coming up with ideas for marketing campaigns, to generating charts and more.

For this blog article, we asked Google’s Bard artificial intelligence chatbot to provide an explanation of how generative AI works. Bard had this to say:

“Generative artificial intelligence (AI) is a type of AI that can produce new content, such as text, images, audio, and video. It does this by learning from a large dataset of existing content. The AI algorithm analyzes the patterns and relationships in the data to identify the underlying rules that govern the content. Once it has learned these rules, it can use them to generate new content that follows the same patterns and relationships.”

AI ROI: How Businesses Benefit from Generative AI

We also asked Bard to list some of the benefits of generative AI. Bard responded that artificial intelligence can increase creativity, improve productivity, and reduce costs. Generative AI, according to Bard, can also help to create new products and services or improve the quality of existing products and services. Bard even cited a McKinsey study showing that generative AI can create 30% more revenue for businesses that use it effectively (though it did not provide a link to said study).

Generative AI can help businesses to better interact with customers or even create new code. There’s no shortage of applications for this type of artificial intelligence; moreover, its list of use cases is growing by the day.

When asked for a list of applications that businesses can use AI for, Google Bard replied that AI can help to automate repetitive tasks like data entry, provide insights by analyzing large amounts of data, and help personalize employee and client experiences.

Here at Appara, our product uses artificial intelligence to auto-generate documents and workflows, highlight errors and omissions, and help maintain accurate records, enabling users to achieve tasks in up to 75% less time.

Limits and Risks: Generative AI is a Tool

Generative AI has a number of strengths and benefits; however, it also presents some limits and risks. One of the risks of generative AI is the possible danger of copyright infringement – namely, copying content from other sources when creating its own answers.

For example: During our testing, we asked Google Bard to define and explain the term “generative AI”, which it did quite successfully (see above). However, Bard didn’t initially list any sources that it used to come up with that explanation.

When explicitly asked to list its sources, Bard replied that it used information from Gartner, Nvidia, TechTarget, WIRED, and GitHub, in addition to its own “knowledge and experience as a large language model”.

Without that list of sources, we would have run the risk of plagiarism in creating our definition of generative AI. When we confronted Bard with the fact that it plagiarized these sources, Bard replied that “I did plagiarize some of the text” and explained that it is “still learning to be more original” in its writing.

Furthermore, when we asked Bard to provide a link to the McKinsey study mentioned above, it replied that it could not show its sources, as it is only a language model.

Generative AI is also dangerous to rely on for highly technical or factual information, as it can sometimes “hallucinate” and produce incorrect results.

Case in point: When we asked Bard if Appara is based in the United Kingdom (which is false), it replied that Appara is based in Mountain View, California (which is also false). We’re actually based in Vancouver, Canada. Not only did Bard list the wrong city and country we’re based in, but it also claimed that we have offices in New York and India, which we do not.

In essence, generative AI is a tool that must be used wisely and in the proper context in order for it to be useful. While it can expedite work, it can also produce errors.

Generative artificial intelligence is clearly a powerful tool, and one that can help your firm to increase its efficiency and reduce errors. While AI cannot fully be relied on to “think” for itself, it’s quite useful for a variety of applications. With major players like Microsoft and Google embedding generative AI into their products, the nature of work will be shaped by this tool for years to come.

How are you using generative AI in your firm? What are some ways you can reduce errors, cut costs, or improve productivity with an AI assistant?