| AI is becoming indispensable as users of various AI technologies become aware of the advantages of using it. AI can perform frequent, high volume and automated tasks without fatigue, can process big data, and gives out accurate predictions in shorter time frames than humans are capable of. This reduces human errors and provides actionable information quicker than manual processes. The predictive ability of Large Language Models (LMM) has become phenomenal, making it possible for AI to learn human language and conventions with increasing accuracy and give satisfactory responses to human queries. However, the shortcomings of chatbots like ChatGPT which are based on LMMs are becoming evident – their performance is based on what they find on the internet and everything in the outputs is fabricated, as the AI’s purpose is to combine what is available on the internet in a plausible form. |
| It is concerning that there are no firmly universally established ethical codes or standards yet, and the adaptation of legislation to protect users and society from AI’s ‘mistakes’ is proceeding much too slowly. There is a limited number of (highly skilled) people who can program the machines, and few people are required to operate them, thus creating insufficient employment in any economy. There is also the long-held fear that robots may someday render humans useless and/or dominate them. The emergence and public release of chatbots based on LLMs have brought this specter closer to reality. |
| This panel session aims to touch on these debates in society and follow up on the two-day NSTF Discussion Forum held in May 2023. |
