Riding The Wave of Generative AI: Tips for Enterprise Leaders

“Generative AI represents the future of innovation, where machines not only automate tasks, but also generate new ideas and solutions.” – ChatGPT

 Yes, ChatGPT gave me the above when I asked for a quote on Generative AI, and I believe it to be true. 

ChatGPT, a natural language processing tool created with artificial intelligence technology called Generative AI, doesn’t need any introduction. It has taken the world by storm since it was announced in November 2022. Its uses range from conversational chatbots to many productivity tools, and more benefits are being discovered daily. 

ChatGPT uses a Large language model (LLM), an Artificial Intelligence (AI) system taught to understand and generate natural language text. It is trained on massive amounts of text data, often of the order of billions of words.   

While using LLMs to create conversational chatbots is a typical use case, there are many others. You could apply such models to computer languages (code). GitHub’s CoPilot, and Hugging Face and ServiceNow’s BigCode, enable the creation of code and completion of partially developed code. The model uses various data, including existing code snippets, and works across programming languages to do this. Both programmers and non-programmers would be able to use it to create computer programs.  

Biotech startups use NVIDIA’s BioNeMo LLM service to understand biomolecular data.  

Content platform Jasper raised $125M at a $1.5B valuation in October 2022, giving credibility to this use case. According to a TechCrunch article, Jasper’s language models, which are “fine-tuned for customer specificity,” “power Jasper’s browser extension for Chrome that delivers contextual content recommendations across platforms including Google Docs, Gmail, Notion and Hubspot.” 

A 2020 paper called Legal Language Modeling with Transformers says, “The law is the domain of natural language, and natural language models could be used for many tasks, such as the generation of contracts, briefs, and rulings. Many legal writing tasks are somewhat repetitive, and in these tasks especially legal language models could save practitioners significant time and resources.” A company called Lexion is doing precisely that. 

Generative AI is not limited to text models. Open-source DALL-E and applications such as AI Magic Tools from Runway offer machine-learning models for images. 

If you are asking yourself, “How can I leverage the Generative AI technology in my business?” there are three things you can do to get started. 

Understand Where the Technology Is 

1. Are you willing to live with a black box solution?
   The system created with LLMs is a black box, meaning you don’t know why the output you see from these models is the way it is. It is like talking to a doctor who wants to prescribe a medicine but cannot explain why you should take it. The vastness of the data these models use makes it impossible to know exactly why the model arrived at the response it gives you. Some use cases require that the results are explainable to the users.
2. Does your solution need to be deterministic?
   A LLM, a machine learning model, is probabilistic, and consistency is not guaranteed. A small change in how to interact with these models could result in a much different outcome. It is acceptable to have a slightly different response when conversing, but if you are relying on it to produce a result that doesn’t change every time you run it, it is a problem.
   The training data is from humans, sourced from various social media, discussion forums, blogs, etc. As such, there is bias in the data. We could create potentially harmful solutions if we don’t use the data with some guardrails.

Identify How Generative AI Can Help Your Business  

1. Does the problem require the use of Generative AI?
   Not all problems require the use of Generative AI. The help desk is one where conversational chatbots can provide much value. Another area where it could help is in generating content for your business blog. It can give you seeds for a blog or create ideas for blog posts, which then can be used by a human writer, making her highly productive. Every business depends on software systems, and it is possible to use tools such as BigCode to make your programmers more productive. However, you need to have experienced programmers guiding the use.
2. Do you have enough data to train a LLM for your business purposes?
   Today ChatGPT uses all the data available on the internet to train the underlying model. However, business applications may require a specific set of data for training. For example, a healthcare application might require medical data that may not be readily available in open-source models.
3. Availability of AI talent
   While tools such as ChatGPT do not require any AI talent to use it out of the box, if you plan to create a proprietary application, you will need to either staff a team that can build the system or consider outsourcing it to a consulting company that has such talent and has a solid reputation in providing AI solutions.

Setup experiments 

Since many are enamored with the shiny new toy, ChatGPT, create a program for your employees to use the tool for anything that will help them be more productive. You can run a competition to see the most beneficial use of ChatGPT, judged by their peers. This could help elevate the understanding of text-based Generative AI. You may also want to explore the image side of Generative AI with open-source tools such as DALL-E. 

Since many in your organization would also question how ready the technology is, you might consider debating its merits and the need to use caution. Be sure to provide a psychologically safe space for such a discussion to take advantage of diverse opinions. 

Many applications are being produced every day which use Generative AI. Research these, and choose a couple to understand what they can do for your business. For example, Addy.ai is an email application that could boost your staff’s productivity by helping them write emails faster and better. 

Create and fund a pilot project to get started if you have identified a proprietary need for LLMs. 

With all of these experiments, be sure to have robust measurements in place to quantify the business benefit. Share these with everyone in the organization to help you scale the efforts broadly in your enterprise. 

Conclusion 

Generative AI has the power to change the enterprise landscape. It is just getting started, and there are still many hurdles to overcome. But the race is on, as evidenced by the competition between Microsoft and Google in the search space. Firms such as Grandview Research predict Generative AI market will be worth over $100 Billion by 2030. The time to begin understanding, researching, and planning to leverage it is now.

Written by Shantha Mohan Ph.D.
Have you read?
Pinduoduo Inc Changes Name to PDD Holdings Inc.
Truth Coaching with Randy Howell Jr. Has the Power to Heal.
Advice for M&A success in 2023 by Josh Rogers.
Ever Heard of ChatGPT for Executives? Better Update Your Resume by Rick Andrade.
5 Opportunities Every Leaders Should Be Embracing in 2023 by Gordon Tredgold.
The Future of Work: Going Beyond the Surface Level Issues by Joshua Berry.