The Potential for Generative AI is Most in Language, Not in Images

Recent interest in generative AI has reached a fever pitch.

Artificial intelligence (AI) that can create new content, as opposed to just evaluating or acting on pre-existing data, is known as generative AI. No subject in the world of technology is currently receiving greater attention and enthusiasm.?

Text-to-image AI has been the blazing hot heart of today's generative AI mania. Based on straightforward written inputs, text-to-image AI models create original, detailed visuals. (For a few illustrations, go here.) Among the most well-known are the Stable Diffusion and Mid Journey models, as well as OpenAI's DALL-E.?

The rapid appearance of these text-to-image AI models during the summer sparked today's generative AI frenzy, complete with billion-dollar fundraising rounds for young firms, extravagant company launch parties, constant media attention, and hordes of entrepreneurs and venture capital firms swiftly rebranding themselves as AI-focused.

Generative AI has come a long way toward being not just quicker and cheaper, but in some circumstances better than what humans develop manually. From social media to gaming to advertising to architecture to coding to graphic design to product design to law to marketing and sales, any field that depends on human creativity is open to disruption. Though it's possible that generative AI will replace certain jobs entirely, others may benefit more from a collaborative, iterative creative process involving both humans and computers. Either way, generative AI has the potential to make the creative process in many different industries more efficient, effective, and affordable. The ideal of generative AI is that it would reduce the marginal cost of creative and intellectual work to zero, unleashing massive increases in both labor productivity and economic value, as well as a corresponding market cap.?

There are billions of people in the information and creative industries that generative AI targets. With the help of generative AI, these workers may improve their productivity and inventiveness by at least 10%. They will be able to do their jobs in less time while producing higher quality work. As a result, Generative AI has the potential to create economic value worth trillions of dollars.

It makes sense that text-to-image AI, unlike any other segment of artificial intelligence, has so piqued the public's attention. Visual content has all the makings of a viral sensation: it's engaging visually, simple to understand and easy to share.?

Additionally, text to image AI is a very potent piece of technology. These models are capable of creating photos that are breath-taking in their originality and sophistication. Advertising, gaming, and filmmaking are just a few of the industries that image-generating AI will change.

The most significant innovation made by humans is language. It is what distinguishes us from every other species in the world more than anything else. Language enables us to reason abstractly, generate sophisticated concepts about what the world is and could be, convey these ideas across generations and geographies, and build upon them. Without language, almost no aspect of modern society would exist.

Every industry, business, and commercial transaction in the world depends on language. Language is essential to the smooth functioning of society and the economy.?

Thus, the ability to automate language opens up hitherto unheard-of possibilities for value creation. The way that every company in every industry operates will change as a result of AI-generated language, as opposed to text-to-image AI, whose effects will be felt most sharply in a few areas.?

Let's go over a few sample applications to show the scope and depth of the upcoming transformation.

From Technology to Business

Marketing-related copywriting, including website copy, social media postings, blog content, and more, has proven to be the first commercially successful application of generative text.?

Over the past year, AI-powered copywriting has experienced astounding revenue growth. One of the top businesses in this field, Jasper, only became operational 18 months ago but is already expected to generate $75 million in sales this year, making it one of the fastest-growing software startups in history. Jasper just disclosed a $125 million fund round, giving the company a $1.5 billion valuation. Unsurprisingly, a plethora of rivals have appeared to compete for this market.?

However, copywriting is just the start.?

Large language models are ready to be used to automate a number of components of the larger marketing and sales stack (LLMs). Expect to see generative AI technologies that, for example, automate outbound emails from sales development representatives (SDRs); appropriately react to inquiries from potential customers about the product via email; handle email contact with clients as they travel through the sales funnel; providing real-time guidance and feedback to human sales agents on conversations; summarize sales dialogues and propose next actions; and more. Human sales professionals will be freed up to concentrate on the distinctively human components of selling, such as client empathy and relationship development, as more of the sales process gets automated.?

Generative AI will significantly automate contract drafting in the legal industry. LLM-powered software solutions will eventually handle a large portion of the back-and-forth among legal departments on deal agreements by understanding each client's unique objectives and preferences and automatically working out the language in transactional agreements in accordance. For businesses of all sizes, post-signing, generative AI technologies will substantially ease contract management.?

Legal research, discovery, and other aspects of the litigation process will be transformed as a result of language models' powerful ability to summarize and respond to questions concerning text documents.?

Medical record writing will be simplified with the help of generative language models. They will answer inquiries regarding a patient's medical background and describe electronic health records. They will assist in automating time-consuming administrative procedures including prior authorizations, insurance claims handling, and revenue cycle management. They will soon be able to provide diagnosis and treatment plans for specific individuals by fusing a thorough knowledge of the body of research literature with the unique biomarkers and symptoms of a certain patient.?

Generative AI will revolutionize customer service and call centers in a variety of sectors, including hotels, e-commerce, healthcare, and banking sectors. The internal IT and HR helpdesks operate similarly. AI and ML can also revolutionize Mobile App Testing!

Many tasks that are performed before, during, and after customer service encounters, such as in-call agent mentoring and post-call documentation and summarizing, can already be automated using language models. They will soon be able to handle the majority of customer service interactions from beginning to end without the need for a human, and not in the stilted, rigid, rules-based manner that automated call centers have operated for years, but rather in fluid natural language that is practically indistinguishable from a human agent.?

Simply put, the vast majority of your future contacts with a brand or organization, regardless of the issue, may and will be automated.

There will be extensive use of automation in the news industry. While human investigative journalists will still seek out stories, generative AI models will increasingly be used to produce the articles themselves. Soon, a large portion of the internet media we access on a regular basis will be produced by AI.?

LLMs will be used by legislators to assist in the drafting of legislation. They will be used by regulators to assist in converting legislation into comprehensive rules and codes. They will be used by bureaucrats at all levels, from the federal to the municipal, to help organise the various administrative state tasks, such as handling permit requests and levying small fines.?

In the academic world, generative language models will be used to create funding proposals, summarize and analyze the body of literature, and, yes, even to create research papers (both by students and professors). There will undoubtedly be a controversy involving students who use generative language technologies to have their essays written for them in class.?

Generative language models will hasten scientific discovery in and of itself. The complete body of knowledge and research that has been published in an area will be assimilated by LLMs, who will then be able to offer solutions and exciting new research possibilities.?

This has already been accomplished; it is not a hypothetical future scenario. Large language models have recently been demonstrated to be able to extract latent information from the body of existing materials science literature and then suggest novel materials for further study by a team of researchers from UC Berkeley and Lawrence Berkeley National Laboratory.

More Than Just Natural Language?

Generative language models (LLMs) have the potential to transform software development, making it one of the most intriguing commercial uses of GLMs.?

Languages, such as Python, Ruby, and Java, are used to create programmes for computers. Programming languages, like English or Swahili, are represented symbolically; each has its own syntax and semantics that are mutually compatible within itself. It follows that the same potent new AI techniques that can acquire astounding familiarity with natural language should also be able to master programming languages.?

We live in a software-driven era. A half trillion dollar market for software is thought to exist today. It's fair to say that software is now essential to running a successful business in the 21st century. As a result, the possibility of automating its production poses a staggeringly massive opportunity.?

Microsoft is the 800-pound gorilla in this market, having pioneered the space. Microsoft introduced Copilot, an AI coding companion tool, earlier this year in collaboration with its subsidiary GitHub and its close partner OpenAI. Codex, a sizable language model from OpenAI, powers Copilot (which in turn is based on GPT-3).?

Soon after, Amazon released CodeWhisperer, its very own AI pair programming tool. Google has also created a comparable tool, however it is not made available to the public and is only used internally.?

The revolutionary potential of these technologies is becoming clear even though they have only been on the market for a short period of time.?

In recent research, Google discovered that its AI code completion tool helped employees code faster and more efficiently than those who did not use it, with the AI writing 3% of their code.?

Even more impressive is new research from GitHub, which discovered that using Copilot can cut the time needed for a software engineer to accomplish a coding task by 55%. Up to 40 percent of GitHub's code is being generated by AI, according to the company's CEO.?

Think of how much more productive Google and Microsoft could be if they implemented something like this company-wide. There is a potential value of billions of dollars at stake.

Three broad points are worthwhile after going over a variety of potential commercial uses for generative language models.?

First, some readers may be wondering whether the use cases presented here are genuinely conceivable, especially those who have not had much experience dealing directly with today's language models. Will generative language models actually be capable of producing a contract, exchanging emails with a potential customer, or drafting a piece of legislation—not just in a strictly controlled demo or research context, but in the midst of all the chaos of the real world??

Yes, it is the answer.?

The great majority of the stuff that humans produce, including the messages we write, the ideas we explain, and the proposals we submit, is not original.?

It may come off as harsh. However, the majority of website material, emails, customer service interactions, and even majority of legislation actually include very little meaningful originality. Despite the differences in word choice, the underlying structure, semantics, and ideas are predictable and constant, reflecting language that has been written or spoken a million times before.?

The large corpora of current material that today's AI has been educated on allow it to learn these basic structures, semantics, and concepts, and when asked, it can convincingly duplicate them with new output.

We will discover that a surprisingly big portion of human language production—those that are essentially non-original—can be effectively automated by LLMs.?

The second broad observation is that each output from a language model can also be used as an input to another language model, which is a key factor in why generative language models will become so strong. This is because text is used as both the input and output modalities in language models. The main distinction between text-to-image models and language models is this. This seemingly insignificant fact has far-reaching ramifications for generative artificial intelligence.

Why is this important? Due to the fact that it permits what has become known as "prompt chaining."?

Despite the fact that huge language models are tremendously powerful, many of the jobs that we will want them to complete—specifically, activities that call for intermediate actions or multi-step reasoning—are too complex to be handled by a single run of the model. Prompt chaining allows users to divide a large task into a series of smaller ones that the language model may complete in turn, with the results of the previous work acting as the input for the next.?

By cleverly chaining together prompts, LLMs are able to perform significantly more complex tasks than they otherwise could. Incorporating a search engine query or a URL pull as a step in a chained sequence of prompts allows models to access data stored in other systems.

In order to deal with generative language models, a new firm called Dust has developed tools that make prompt chaining clear and easy to understand. When a user types in a question like "Why was the Suez Canal blocked in March 2021?" Dust will fetch the top three results from Google, read the content on those sites, summarize it, and then synthesize a final answer that includes citations.

Incorporating prompt chaining into the development of LLM-powered apps will make it simpler to design modular, extendable, and interpretable software. It will make it possible to create sophisticated software applications with broad functionality. This recursive richness is unmatched by anything in text-to-image AI.?

This gets us to our third and last point: one of the most essential factors to consider when commercializing and implementing LLMs will be how and when to include a human in the loop.?

The majority of generative language applications won't be initially deployed entirely automatically. It will always be wise or required to have some amount of human control over their outputs. Depending on the application, this will take on a variety of different appearances.

Iterative and collaborative interaction will soon be the most natural way for humans to utilize LLM applications; in other words, the end user will be the human in the loop. For example, the human user might give the model an initial prompt (or prompt chain) to produce a specific output, evaluate the output and then modify the prompt to enhance the output, run the model numerous times on the same prompt to choose the most pertinent versions of the model's output, and then manually modify this result before launching the dialect for its destined use.