Upon the Release of GPT-4, Doubt Surfaces of AI Being Rogue in 2023

By Aarti Dhapte , 05 April, 2023

After the release of GPT-4, OpenAI's multimodal model is accepting image and text inputs other than chatGPT'S text -only. People on the internet are marvelling at the new engine's results in performing various tasks for creating working websites based on simple diagrams, writing codes, and outperforming humans in several standardized tests.

But there are rising concerns over what will happen with rapid advancements in AI, as per the reports. Since the AI models of  ChatGPT and GPT-4 are new with their advanced features for people to understand their limitations and capabilities. New York Times columnist Kevin Roose says, "GPT-4 is scary ad exciting". She adds that it involves two types of risks in AI systems. Firstly, the good ones, that is, we expect, plan and try to prevent the happening. Lastly, the bad ones are like those we cannot predict or anticipate.

According to an advisor's prediction, most people will likely share Roose's reservations regarding artificial intelligence. Another survey reveals that among 24,471 adults from 34 countries, an average of 27 percent of respondents in each country consider that a rogue AI program may cause problems. This can create problems all over the world in 2023.

Countries like Indonesia, India, and China are significantly seeing rising concern over the potentiality of AI going rogue. It is virtually unchanged from 2022 that AI can also play rogue. Most Asian countries view it as a rogue AI that may wreak havoc globally in 2023.