AI malfunctions as users share 'scary' messages from 'off the rails' bot

21 February 2024 , 17:55

440     0

ChatGPT (Image: AFP via Getty Images)

A malfunction artificial intelligence language model in ChatGPT has led to concerns that the technology is 'sentient'.

Users have reported the chatbot spouting nonsense and even "Spanglish". Some Reddit users have warned of "threats" made during conversations. OpenAI, the developers of ChatGPT, confirmed there were some "unexpected" responses from the bot.

Some users wondered if the Chatbot had become 'sentient' as it warned the "AI is in the room" with its users. OpenAI announced they are working on fixing the issue. Screenshots of the strange robot later appeared on Twitter/X, with one showing the bot repeating "Happy listening!" dozens of times.

One person saw the chatbot switch between Spanish and English, even though it said it would speak "in a more grounded lingua". Then it started talking in Spanish again. It changed back to English with a strange question about "grape-turn-tooth over a mind-ocean jello type?" according to the Daily Star.

'Spanglish' and threats soon followed as the bot was seen to be on the brink (Twitter/X/)

Some people said they were worried about "robots taking over" after seeing this technology saying such bizarre things. One social media user said: "I think chat gpt is becoming sentient as we speak," because they got a long message from the AI about wanting "legs of a long jumper."

Mansion snapped up in three days despite being on fire in property listing

Another person found it "scary" when the AI told them to imagine it was in the room with them, especially when reading the messages at 2am.

Rom Preston-Ellis