This July, two of the top most trending news contains both these hashtags, “Facebook” and “Artificial Intelligence”. Before going to know about the rumor- “Facebook shutting down AI”, let’s examine the opinion clashes between both Facebook CEO and Tesla CEO regarding AI’s performance advancement.
Past this month at US National Governors Association annual meeting Elon Musk, Tesla CEO/Space X Founder reacted on the regulation of AI: “AI is the rare case where I think we need to be proactive in regulation instead of reactive,” since he foresight the Horizon’s danger. he also added, “AI is a fundamental existential risk for human civilization…but until people see robots going down the street killing people, they don’t know how to react.”
And few days after that, Mark Zuckerberg- Facebook CEO, responded to Musk’s statement on AI in a Facebook Live post stating: “I think people who are naysayers and try to drum up these doomsday scenarios—I just, I don’t understand it,” emphasizing the enormous benefits of AI, he added “It’s really negative and in some ways I actually think it is pretty irresponsible.”
To that, Musk the very next day tweeted: “I’ve talked to Mark about this. His understanding of the subject is limited.”
This is not the first time Musk responding in this way to the topic of AI. Even two years back, when he was interviewed at MIT AeroAstro Centennial Symposium, he cautioned about taking care in the regulation of AI Development mentioning that it is our “biggest existential threat” and compared building it (AI) to “summoning the Demon.”
#If we take a closer look on both these Game changers’/Billionaires’ opinions on this vast and unpredictable AI:
What Musk senses about the danger by the uncontrollable AI’s advancement should definitely hold some merit. The fact that computing power doubles every two years, Moore’s law never went wrong in the past 50 years. So, since anything advancing that quickly demands our attention, regulating AI becomes a problem afterward if not now.
Now, after Mark’s views over AI got crystal clear, social media got shaken with one whole new twisted news which got viral overnight: “Facebook shuts down AI after it invents it’s own creepy language.” claiming that concerned AI researchers of Facebook abandoned an experimental chatbot program after realizing that AI bot agents were inventing their own language.
Did Facebook really shut down AI after the bots inventing their own language?
The idea is actually to build up the negotiation between “AI-to-Human” to make their users feel easier in doing an AI chatbot interaction. For that sole purpose, the team taught their agents to talk. For that, they fed the computers with hundreds of thousands of dialogues between humans to make them learn by themselves from trial and error technique called “Reinforcement Learning”.
Media raised the news in such a way that Facebook AI Researchers in panic, got forced to kill their creepy bots which started speaking in their own language. But that’s not quite what happened. Facebook’s researchers observed that the bot agents learned to use tricks of repetition to communicate what they want. so, convo below cannot be crossed away easily by saying it as some weird bantering.
Bob: i can i i everything else . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to
Still, since the researchers hoped to make bots that could converse with humans, they are still redesigning the training schemes to make sure the bots learn some recognizable language because although computers make decisions in a way that we may not understand, it still does use the information we feed it.
Reply from Facebook AI Research Team to the media:
And to this errant media coverage, Dhruv Batra of Facebook’s Artificial Intelligence Research group clearly explained what was the output meant. He said “Like if I say ‘the’ five times, you interpret that to mean I want five copies of this item. This isn’t so different from the way communities of humans create short-hands.” He also added his notion to this false news spread, “While the idea of AI agents inventing their own language may sound alarming/unexpected to people outside the field, it is a well-established sub-field of AI, with publications dating back decades.” clarifying the silliness in the news, “changing the parameters of an experiment is NOT the same as “unplugging” or “shutting down AI.” If that were the case, every AI researcher has been “shutting down AI” every time they kill a job on a machine.”
Why these clashes and collisions in opinions about AI?
The current application of AI, “Machine Learning“, is not like other regular hard-core handmade programming where each scenario gets detailed instructions in what way it should respond.
Machine Learning means it gives the machine lots of information which then gets to learn on its own and use that info, to respond according to its own knowledge.
We need to know one thing here. till now, machines/ computers used to have limited capacity to do what they were asked to do. Humans maintained the control. But as we are handing over more and more decision making opportunities/ processes to the algorithms and automation which we ourselves don’t know what exactly could happen in future, there is a strong need for ourselves for regulating the advancement of AI proactively.
So, there you go, Facebook didn’t “shut down” AI. Thank you so much for reading our blog. Please Let us know what you think about AI and whom you vote between Musk and Mark.
Well, as you seem to be quite interested in AI, to know more about what’s happening around the world, around AI or around your own business field, Log on to SmatSocial and get introduced to our social media AI assistant.