Democratisation of AI

24-04-2023

Opinion Article: "GPT: After the hype and fright storm – will this be the democratisation of AI?"

Opinion Article Paulo Noorm (2023)_Inside News_1400x700px

The latest stream of news and announcements concerning the new variables of generative Artificial Intelligence (AI) has brought a storm of overstatements and fears concerning the topic. OpenAI's ChatGPT seemed to be becoming a part of society, and now, since the release of the newer version, GPT-4, we have witnessed a leap towards the democratisation of AI: with enough generative text to write a book, the ability to code in all programming languages and, most impressively, image decoding.

 

First, we felt the "fright". Opinions that these new tools (LLMS) could replace a diversity of professions overnight started to come up. Nevertheless, we can relax, robots are not yet ready to replace human labour completely. However, language models, such as ChatGPT, could act as an accelerator - operating at high velocity on a large scale - with the potential to  benefit the labour market and society. Essentially, the more we feed these models with data and questions, the faster they "learn" to predict results. Currently, some start-ups claim to use these tools to create new products that can revolutionise society, from legal administration to stock trading, games and medical diagnostics, even though the reality is that curiosity has mainly led some to seek ChatGPT to help them with questions such as writing poems or to test a system out to its error.

 

Today with these new platforms, we have started to feel a democratisation of access to Artificial Intelligence across society. If not long ago, the implementation and execution of AI systems were only within reach of a select few, they are now at a distance of a click for registering in a platform, whether for personal use or work-related. Conscious of the importance of taking the lead in this growing tendency, other companies, such as Google, are quick to mark their presence and are already preparing the launch of platforms to compete with ChatGPT.

 

Such free platforms enable the creation of much more convincing phishing e-mails and the script of more advanced malware, meaning they are not risk-free. It is not surprising some financial institutions are not open to testing this technology, at least for now. Promoting a customer-centric experience based on transparency and trust is crucial for industries where confidentiality is paramount. In sectors such as banking, it is essential to invest in 100% reliable technology providers who can offer high-tech solutions that protect their customers from these threats.

 

I believe it is still early to give these new platforms too much credit, for they still need to evolve to start playing a dominant role in society.  The influence of all forms of AI on the ethical, economic, political, and social spheres should be discussed in light of these recent technologies, and software developers need to be prepared for any new cybersecurity risks they may provide. 

 

News: Dinheiro Vivo