ChatGPT has good answers for almost any topic. You can get him to write a song for you, or you can give him his five-part framework for your company’s digital strategy. For most common topics, like this example, the output will probably be sensible. But in a more specific question, you may be missing quite a few details.
People have used generative AI to negotiate discounts on phone bills, treat real patients, write Python code, poems, songs, novels, take exams (or even cheat). I was. In general, Large Language Models (LLM) produce good results that look great.
Therefore, they can indicate changes in how communication and business work. But it’s too easy to think it’s time to give room to AI champions. Several authors have written with irony about how AI will put companies out of business. Such panic is wrong. To understand its potential, let’s take a look at how AI tools like ChatGPT work, what they can do, and how businesses can use them.
What’s behind the interface?
The latest generation of AI is based on LLM. Interestingly, ChatGPT combines LLM with a dialogue layer that uses reinforcement learning.
LLM is a neural network model that uses unsupervised learning to predict outcomes. Among the many AI models developed, LLM is uniquely difficult to explain.
Language models (unlike large language models) have existed for a long time and can predict the next word or phrase in a sentence. It uses a different technology than LLM and is used for different purposes. Autocorrection is commonly used.
Adding the “large” factor involves training the model on a large collection of publicly accessible electronic documents. That collection (a “corpus” in AI parlance) consists of several petabytes of data (one petabyte equals one million gigabytes). Training a model on such a large amount of data will allow it to learn not only language patterns, but many topics as well.
So LLM is “big” in part because of the amount of data used for training. But the size of the model itself also matters. A few years ago, a complex model might have hundreds of parameters. LLMs have billions of dollars in assets. ChatGPT’s underlying LLM has 175 billion parameters and has been trained on about 500 billion “tokens” (words or word fragments).
The progress we’ve seen so far is largely the result of our efforts to answer one question. So how can a model with so many parameters do anything useful?
AI is also being integrated into business operations in the Philippines, using these advanced AI technologies for a wide range of applications such as natural language processing, predictive analytics, and machine learning.
However, it is important to note that the use of AI also raises ethical and privacy issues, especially regarding the use of personal data. Therefore, Jalan Marcel S. Manrique, Head of Technology Consulting at KPMG in the Philippines, said, “When companies exercise power, they should adopt a mindset that puts transparency, responsibility and ethics at the core of their decision-making process. is the most important,” he suggests. of artificial intelligence.
“Doing so will help companies navigate the complexities and potential pitfalls associated with AI adoption, while fostering an environment of trust and accountability among stakeholders,” he added.
This excerpt is taken from a KPMG Thought Leadership publication: https://kpmg.com/xx/en/home/insights/2023/02/the-potential-impact-of-chatgpt-and-the- new-ai-on-business.html.
© 2023 KPMG Intl Ltd. is a privately held company with guarantees in the UK and has a Philippine partnership RG Manabat & Co. as a member company. all rights reserved.
For more information, please email email@example.com or visit www.home.kpmg/ph.
This article is for general information purposes only and should not be considered professional advice for any particular issue or group. The views and opinions expressed herein are those of the author and do not necessarily represent his BusinessMirror, KPMG International, or his KPMG in the Philippines. Please email firstname.lastname@example.org.