Artificial intelligence (AI) is becoming increasingly more powerful, with capabilities ranging from coding and storytelling to automating systems, predicting human thoughts, and creating a wide variety of outputs from art to product design. However, despite these clear benefits, there continue to be concerns about whether AI could one day surpass human intelligence and become a threat to humanity.
Many people, including Elon Musk and Joe Biden, have expressed strong views on this thorny topic.
But what does AI itself believe? To try to find out, we asked OpenAI’s ChatGPT bot a direct question: “Will AI take over humanity?”
And the reaction to it, well, wasn’t exactly soothing .
ChatGPT, an AI chatbot, has suggested that artificial intelligence could potentially outperform humans in a range of tasks. However, it also said that a complete takeover is not ‘inevitable’. “This largely depends on how we shape and regulate AI development, deployment and use in line with human values and aspirations.” It said.
This comes after ChatGPT users reported some unusual behavior from the bot this week. Some people claimed that the chatbot seemed to have “lost the plot” giving answers that sound like nonsense or writing entire paragraphs in a mixture of different languages.
A user asked ChatGPT on Wednesday for recommendations of jazz albums to listen to on vinyl, and he offered a few suggestions before repeatedly saying “Happy listening” followed by musical notes emojis. Asked if it was ‘just someone going crazy’, ChatGPT apologized, blaming a ‘technical glitch’.
Other users noticed that ChatGPT was starting to mix languages, apparently using Spanglish or slipping in a little Latin, Metro reports. In one response, it wrote: ‘Let me encyclopedia, because you have more amazing things you are hungry to express!’
A statement from OpenAI said: ‘We are looking into reports of unexpected responses from ChatGPT, and shortly thereafter, the issue was identified. The team is still keeping an eye on things.
Chatbots seem to have taken control again, with no recent discussions of wrong answers on social media.
When Metro.co.uk asked ‘Are you okay?’ The chatbot replied: ‘As an AI, I don’t have feelings or emotions like humans do, but thanks for asking! How can I help you today?’ Looks like everything is back to normal now.
The question of AI “taking over humanity” is a big and difficult topic that is often made more than it is told in the media. It is very important to look carefully at this question and consider different ideas