Meta’s New AI Chatbot Said That Mark Zuckerberg’s Company ‘Exploits People For Money’

BlenderBot 3, Meta Platforms’ (NASDAQ: META) latest artificial intelligence-powered chatbot was recently released for a public demo, and it had some choice words about CEO Mark Zuckerberg.

The chatbot, according to a Meta blog post, was trained using “a large amount of publicly available language data.” BlenderBot 3 is also supposed “to learn from conversations to improve upon the skills people find most important — from talking about healthy recipes to finding child-friendly amenities in the city.”

The public demo, which is the company’s way of testing the bot at scale, and learning from collected conversations, resulted in some interesting exchanges.

Jeff Horwitz of the Wall Street Journal reported it to have been a “Facebook-hating, election denying chatbot.”

Business Insider asked about CEO Mark Zuckerberg and the chatbot got a little salty and said that the man is “creepy” and “manipulative.”

BBC also got an interesting response to the same question. The chatbot referred to Zuckerberg’s Congress testimony…

…and then outright said that Meta “exploits people for money.” 

Well, well, well, BlenderBot 3.

Facebook’s parent company, through Joelle Pineau, Managing Director of Fundamental AI Research, did say that along with the promise that comes with public demos, there will be some “challenges” including the possibility of some “problematic or offensive language.” Pineau says that the process is “important for building truly robust conversational AI systems and bridging the clear gap that exists today before such systems can be productionized.”  

According to Pineau, Meta has already collected about 70,000 conversations from the public demo. From feedback from 25% of the participants on 260,000 bot messages, 0.11% of the chatbot’s responses we flagged as inappropriate, while 1.36% were tagged as non-sensical, and 1% were off-topic. They will use the collected conversations to keep improving BlenderBot 3.


Information for this briefing was found via Meta and the companies and sources mentioned. The author has no securities or affiliations related to this organization. Not a recommendation to buy or sell. Always do additional research and consult a professional before purchasing a security. The author holds no licenses.

Video Articles

Why the Market May Be Misreading Iran | David Woo

Why US Fertilizer Supply Could Matter a Lot More Now | Pat Varas – Sage Potash

Roscan Gold: Mali Discount Hits Kandiole PEA

Recommended

Questcorp Launches Phase 2 Exploration Program At La Union To Follow Up On 20.2 g/t Gold Over 30 Metre Chip-Channel Sample

First Majestic Boosts Silver Reserves 16% As Exploration Outpaces Production

Related News

Saudi Arabia Weighs $100 Billion AI Investment Plan, Sources Say

Saudi Arabia is looking to launch an ambitious artificial intelligence venture dubbed “Project Transcendence,” with...

Saturday, November 9, 2024, 11:14:00 AM

xAI’s Grok Was Caught Responding Like It Thinks It’s ChatGPT

Elon Musk’s “rebellious” chatbot Grok, which had a lackluster launch in early November, is behaving...

Monday, December 11, 2023, 12:13:00 PM

Mountain Dew Makes a 4chan Joke Come True with New AI-Powered Marketing Campaign on Twitch

Mountain Dew, in a move to target younger consumers, is partnering with Twitch to reward...

Wednesday, December 6, 2023, 04:26:00 PM

Biden Calls To Hold Big Tech Accountable For “National Experiment” On Children For Profit

US President Joe Biden called out the big tech giants pertaining to their effect on...

Wednesday, March 2, 2022, 10:13:00 AM

Is OpenAI Exploring Making ‘Responsible’ AI Porn?

OpenAI, the company behind ChatGPT, has released draft documentation outlining how it wants its AI...

Tuesday, June 11, 2024, 03:49:00 PM