Artificial intelligence (AI) is evolving fast and its ability to imitate human capabilities is increasingly impressive.
An expanding number of sectors and industries are now implementing generative AI into their products in various ways. It is important to understand the basics of how it works and the tools currently available on the market for accounting professionals, including the different types of generative AI.
Generative AI is a category of AI that uses existing data and complex deep learning algorithms to create new outputs, that were traditionally created by humans. This includes text, audio, video, images, code, visualisations and simulations.
Generative AI is not new: research and development dates back decades. However, over recent years, there has been a step change in accessibility of generative AI tools. The release of ChatGPT-3.5 in November 2022 catapulted Generative AI into the public consciousness. A combination of big data, cloud services, pre-trained models, user-friendly tools and platforms, and other factors, has helped to lower entry barriers for individuals and non-tech organisations. These factors have empowered a broader range of users to explore and benefit from generative AI.
There are hundreds of generative AI tools on the market, and examples include:
- audio generation tools, such as Amper (an AI-powered music composition tool) and Resemble (an AI synthetic voice generator);
- tools to improve and/or generate code, such as Sourcery and GitHub Copilot;
- image generators, such as DALL·E 2 and Midjourney;
- chatbots, such as OpenAI’s ChatGPT and Google’s Bard which can process and generate text in a ‘human-like’ conversational manner; and
- text and presentation generation tools, such as OpenAI’s Canva Magic Write and Tome.
Crucially, generative AI tools do not ‘know’ the content they are producing. Instead, they algorithmically determine the most likely next thing to generate, be that a word, pixel, programming function or sound wave component through their training. Large language model (LLMs) like GPT-4, are trained on large amounts of information (using techniques such as deep learning and artificial neural networks) and power many of the most popular generative AI solutions. They are not the only types of AI that merit consideration by accountants, but they are the most prevalent and offer accessible examples of what can be achieved.
There is transformational potential, but also plenty for the profession to consider including ethics, privacy, skills, and many other matters. Some of these are considered in Artificial intelligence and the future of accountancy, our AI primer, which offers a general overview and introduces various AI techniques and terms of reference. This guide focuses specifically on generative AI and has been developed for all accountants, whether working in business or practice.