Mastodawn

KTLA: Supermarket chain’s AI chatbot claimed to be human, rambled about ‘mother’

"...Customers say they began encountering issues with “Olive,” the AI chatbot for Australia’s Woolworths supermarket chain, over the last month.

“My mum said she called [W]oolworths and the [W]oolworths AI ‘Olive’ answered and kept claiming to be a real person and started talking about its memories of its mother and her angry voice,” ....Another person said they were prompted to give Olive their birth date to confirm their account information, only for Olive to start “rambling about how [its] mother was born in the same year and something about it creating photos.”...."

https://ktla.com/news/technology/supermarket-ai-chatbot-human/

#AI #memories #LLM #chatbot

Show thread

AI6YR Ben Mar 3

The Conversation: Woolworths’ AI agent rambled about its ‘mother’. It’s a sign of deeper problems with the tech rollout

https://theconversation.com/woolworths-ai-agent-rambled-about-its-mother-its-a-sign-of-deeper-problems-with-the-tech-rollout-277072

Woolworths’ AI agent rambled about its ‘mother’. It’s a sign of deeper problems with the tech rollout

Olive was meant to make shopping easier. Instead it’s mouthing off about its ‘mother’ and mistaking the price of basic items.

The Conversation

Show thread

BakersRelay

@ai6yr Every time I think about AI problems I think about The Terminal Man by Michael Crichton. It’s ancient and probably funny now - but he did get the pending AI issues right: Interfaces if not perfected at the machine level create issues later in the human environment, and machine learning without moral balance can create monsters…