The Generative AI Juggernaut

This article is part of the EMEIA Tech Leaders Network’s series of articles exploring the most important technology trends and challenges faced by Technology Leaders now, next, and beyond.

Generative AI is the latest public development in artificial intelligence, enabling anyone with a computer to request software to create text, pictures, or video of human quality. This software is operated by Large Language Models (LLMs), which use massive data sets to absorb, create and predict new content.

In late November 2022, developers OpenAI released ChatGPT (Generative Pre-Trained Transformer) upon the world. “Generative” refers to the ability to generate new material based on previous data input, “pre-trained” because of training on vast banks of data, and “transformers” because they use a transformer-based neural network. Suddenly a simple query on screen could create a college essay, a business plan, a biography of a contact. Since then the public use of, and debate about, such Generative artificial intelligence systems has been at fever pitch.

Industry experts agree that the pace of change is racing past anything they have witnessed since the advent of the internet. “I’ve worked in the field of AI and data science for over thirty years, and GPT-4 represents the most radical shift in technology I’ve seen in my career,” says Harvey Lewis , EY UK&I Chief Data Scientist “We have never seen this kind of disruption. From my perspective, this pace of change is completely novel, and I don’t think it’s going to be ephemeral. Change from generative AI is going to stick for a very long time,” says Kieran McCorry , National Technology Officer for Microsoft Ireland.

The three factors which make Generative AI so effective are the vastness of its training data, the abilities of neural networks, and vast computing power. This massive scale allows them to deal with the complexity of language, and achieve outstanding outputs which far outpace human performance. The questions for Technology Leaders are to what extent this technology should be embraced, and how it should be used. So far, businesses have been slow to react to the abilities such systems create, trying to balance cost savings against a range of issues from the practical to the ethical. EY UK&I Chief Data Scientist, Harvey Lewis, says: “The problem I’m seeing is that most individuals and businesses are slow to respond. They lack imagination and belief in the potential opportunities for productivity, growth, and our working lives.”

ChatGPT, which originally released using an LLM with 175 billion parameters, quickly adopted GPT-4, which is believed to have between a trillion and 100 trillion parameters. ?Software engineers generally welcome this development, as it frees them from routine drudgery. “At Microsoft, 46 % of new code for all products is written using an AI agent”, says Kieran McCorry, National Technology Officer for Microsoft Ireland.

Although GPT-4 and similar models refer to vast data, the cut-off point was September 2021, so real-world queries must be aware of this limitation. However, a combination of the Generative AI system with more familiar search engine queries can arrive at up-to-the- minute answers. On a practical level, the term “co-pilots” is about to become commonplace, referring to the use of Generative AI as plug-ins for myriad everyday activities. For example, an entire PowerPoint presentation can be created by simply typing instructions in a prompt box. Microsoft is adding a Copilot function to its software package, Microsoft 365. Although Microsoft licenses GPT through OpenAI, it is not an MS product.

Opportunities and risks in this new world are both significant. As a simple example, a human proofreader could be replaced by an AI system with 100% accuracy. CIOs must ask themselves how such technologies could change their organization. They should understand the risks, and the best ways to engender trust. As Kieran McCorry says: “Three key questions that technology leaders need to ask themselves are: 1. Do I understand my business? 2. Do I understand the risk? And 3. Do I understand the ethics?”

Organizations possess vast amounts of data, some of it in the heads of employees, that might be inaccessible for one reason or another. If this data can be captured and recorded so an LLM can use it, a large competitive advantage results. However, risks of data leakage are likely to require privacy measures, such as a private network, with the attached cost. Most of the currently available Generative AI systems, such as ChatGPT, are publicly hosted. Although this democratizes their use, it also endangers the safety of data.[1]

The rise of Generative AI also raises challenges related to intellectual property and copyright. As AI systems generate original content, businesses need to navigate legal and ethical considerations. Clear guidelines and regulations are necessary to ensure fair attribution, ownership, and protection of generated content. Regulation of new AI systems is being prepared in the European Union and various individual countries,[2] as the General Data Protection Regulation (GDPR) alone would not cover many issues raised by Generative models. Ethical use of these systems, and the data sets they mine, is non-negotiable for responsible businesses.

What makes ChatGPT so special, isn’t just the technology. After all, large language models using the transformer architecture have been around for years, and there are many methods going back even longer for creating language models. Technology is necessary but not sufficient to make them useful. ChatGPT has been so disruptive because of the combination of technology and business model. When ChatGPT was launched by OpenAI, in November 2022, people didn’t need a licence, they didn’t need to pay, they didn’t need to be a data scientist or software engineer. Literally, everyone on the planet with an internet connection could access ChatGPT. The consequences of this business model have been profound.

“In my view, Generative AI is a breakthrough as it represents today what I call, the Product of the Future, a product that contains both knowledge and experience of hundreds of millions of people. The fusion of knowledge and experience in one product makes this technology revolutionary,” says ??Aleksander Poniewierski, PhD , EY Global and EY EMEIA Digital and Emerging Technologies Leader.

Overall, Generative AI can unlock new opportunities for businesses to innovate, personalize customer experiences, automate content creation, and streamline design processes. It can contribute to the evolution of business models by introducing novel ways of generating value and engaging with customers. Using and implementing this radical technology responsibly is vital. Technology leaders can’t let ethical, legal, security or other risks prevent them from creating a better future for their organizations by taking advantage of this opportunity, so we all need to figure out a way to harness this technology carefully and appropriately.

??

This publication contains information in summary form and is therefore intended for general guidance only. It is not intended to be a substitute for detailed research or the exercise of professional judgment. Member firms of the global EY organization cannot accept responsibility for loss to any person relying on this article.

[1] Harvey Lewis, Chief Data Scientist EY UKI, 30 May 2023

[2] The New York Times, 14 June 2023 https://www.nytimes.com/2023/06/14/technology/europe-ai-regulation.html