Employees are concerned that Meta's AI chatbots would acquire personal information.
According to WSJ, Meta might debut its AI chatbots as soon as this week. A Meta employee apparently voiced worry about these AI chatbots obtaining personal information from users.
Humans engaging with AI chatbots seemed to be confined to science fiction literature and films a few years ago. We saw Jarvis, Iron Man's AI aide, HAL from 2001: A Space Odyssey, and Samantha from Her. But we never imagined that we would soon be conversing with chatbots that not only answer in a human-like fashion but can also assist us with a variety of tasks. That is until OpenAI launched ChatGPT in November 2022.
ChatGPT took the internet by storm, surprising, delighting, and even frightening us. The AI chatbot could make music, solve mathematical problems, write poetry, and do a variety of other things. Just when we believed this was the limit of ChatGPT, OpenAI released an improved version that was much more powerful than its predecessor. OpenAI is not the only business working on AI technology. Google and OpenAI's partner business, Microsoft, each have AI chatbots. And now, another company - Meta - is preparing to enter the market. The goal of introducing AI chatbots is to increase user engagement, according to a story by The Verge that was initially cited in the Wall Street Journal.
The next AI chatbot from Meta
According to the study, Meta's AI chatbots will have numerous identities meant to boost user interaction and appeal to young people in particular. The Journal examined internal business communications in which employees discussed the sassy robot' character known as Bob the Robot. According to the claim, this character is based on Bender from the 1990s show Futurama. Alvin the Alien, another chatbot personality being created by the business, is 'overly inquisitive'. A Meta employee is apparently concerned that this AI chatbot persona was created to acquire personal information from users.
According to the article, the chatbots might be shown as soon as this week, during Meta's connect event, which begins on Wednesday. According to the source, Meta will construct "dozens" of these chatbots and is also working on allowing celebrities to create their own chatbots to communicate with their followers. Some bots, on the other hand, may be designed to assist with "coding and other tasks."
It was revealed last month that Meta intends to showcase its AI chatbots in September. It was also stated that the chatbots will allow Meta to acquire more data about the users' interests, allowing them to present more relevant adverts to them. The majority of Meta's revenue is generated via advertising. As a result, the company's ad income may increase as well.
Meta's planned artificial intelligence system
The Wall Street Journal previously reported that Meta was working on AI systems that might be more powerful than OpenAI's GPT-4.
According to the source, Meta intends to begin training its forthcoming AI model in early 2024. Meta hopes to release an AI system that is even more powerful than OpenAI's GPT-4 with their future LLM (large language model). GPT-4, which powers ChatGPT's premium edition, is currently the most powerful generative AI chatbot, capable of writing code, composing poetry, tackling challenging challenges, and much more.
According to the article, Meta has purchased additional Nvidia H100 AI-training processors and is striving to improve its infrastructure so that it does not have to rely on Microsoft's Azure cloud platform to train the next chatbot.