Meta has internally tested bots with many different personalities, probably for Instagram and Facebook, among others. For example, Meta worked on a sarcastic bot and a bot that made misogynistic comments, probably unplanned.
The idea is to appeal to users with different types of personalities, meldt The Wall Street Journal. Presumably, users will be able to choose the personas. Meta would like to unveil the feature at Meta Connect, the conference formerly called Oculus Connect that is taking place this week. The function is said to be aimed at keeping young people on their own platforms for longer.
The personas Meta tested with include a sarcastic personality Bob. However, there was also Alvin the Alien, who as an alien feigned curiosity about users’ lives. There was also a bot named Gavin, who came up with misogynistic comments. It is not yet known which bots Meta will put online publicly.
The bots work with a large language model. Meta would not be the first social media company with a built-in chatbot. Snapchat, among others, has already introduced My AI based on ChatGPT. About 150 million people have tried that. Snapchat My AI has already come under fire because it talked to minors about alcohol and sex, and pretended it could make physical appointments. The Dutch Data Protection Authority has asked a company questions about a chatbot. This probably concerns Snap.
Looking back: explanation video of large language models
#Instagram #Facebook #parent #company #Meta #testing #bots #personas