To make sure you never miss out on your favourite NEW stories, we're happy to send you some reminders
Click 'OK' then 'Allow' to enable notifications
Featured Image Credit: The Photo Access/Alamy Live News. Alexey Kotelnikov / Alamy Stock Photo
According to Meta, the new system was created so that humans could have longer, more in-depth conversations with AIs.
It was developed using a ‘large amount of publicly available language data’, consisting of more than 20,000 conversations over 1,000 topics.
And while BlenderBot 3 only launched last Friday (August 5), it wasted no time bashing the head honcho.
The bot added: "He's too creepy and manipulative."
Talk about a system failure.
BuzzFeed’s Max Woolf also revealed via Twitter that the new system also went for the jugular and criticised Zuckerberg’s fashion sense.
The bot said: “He is a good businessman, but his business practices are not always ethical. It is funny that he has all this money and still wears the same clothes!”
Others also shared via social media that the AI was spitting fire about Zuckerberg on their end too.
One person reported that the bot said of the 38-year-old: “He is a great guy but I’m not sure he should be in charge of such a big company.”
Before adding: “He seems to care more about his image than the company itself. But maybe that will change with time.”
Another person revealed that when asking if the company has spoken to Zuckerberg about paying its content creators more fairly, the AI replied: “No I haven’t had the chance yet but hopefully soon after this interview we can sit down and talk about how much money Facebook makes from our content.”
According to Meta, while they’ve conducted large-scale studies and developed new techniques to ensure the bot doesn’t mimic unsafe, harmful remarks, there still might be glitches.
With this prototype, they are still collating feedback to improve the device, which I imagine will be a field day.
A spokesperson for the company also said: "Everyone who uses Blender Bot is required to acknowledge they understand it's for research and entertainment purposes only, that it can make untrue or offensive statements, and that they agree to not intentionally trigger the bot to make offensive statements.”