This article is produced by NetEase Smart Studio (public number smartman 163). Focus on AI and read the next big era!
In the past few weeks, you may have seen your Facebook friends share an article about how a pair of AI-driven Facebook chat bots deviated from their initial setup and invented their own. linguistic.
The key is: Facebook created two artificial intelligence chat programs that talk to each other (and learn from each other). Eventually they stop communicating in English and start communicating in a language other than English. These languages ​​are “invented†by themselves. ".
The front page of the news has reported on this incident, from the eye-catching gimmicks to the analysis of twitches, such as the Daily Telegraph's "Chat Robot invents a new language to communicate Facebook forced to choose to close," making people sound like these chats. Robots are plotting to fight against humanity or pose a threat to humans. Take a look at these comments under the news, and you will find that everyone's reaction is frightened, excited and even happy. Some people say "this day has finally come" or quote from other Isaac Asimov novels.
But just as Etirouft Batra of the Facebook Artificial Intelligence Division pointed out on Monday, artificial intelligence has been “inventing†new ways of communication for decades, so this is not really news. In addition, he explained that contrary to the news headlines, this experiment was not closed down and was changed only to adjust language exchanges.
In addition, the company did not try to hide experimental details. Everything is public, all the details can be found on GitHub, allowing other programmers to copy this scene.
The bottom line is: In terms of artificial intelligence, the new artificial intelligence language should not be a problem we worry about.
Here's what we should worry about
Technology-savvy people like Elon Musk and Bill Gates have contributed a lot of time and money to explaining that artificial intelligence is a threat and looking for new ways to improve their moral standards.
These are the main areas that we should focus on:
arms:
Armed military drones have been put into use, and automated weapons are considered the third war revolution. Automated smart weapons have caused fewer soldiers to be harmed and are far cheaper than advanced technologies such as fighter jets or nuclear materials. For example, current "Reaper" drones cost about $13 million, while a fighter jet costs $100 million. In the future, they will become cheaper, and unregulated (and extremely powerful) mobile weapons will also be cheaper, and anyone can programmatically kill and destroy them (unless we take the initiative to take measures
Shi).
control:
Perhaps the most important thing in control is not the artificial intelligence itself, but who controls it. We are dealing with those who have the ability to reshape the world. If they are monopolized by greedy companies, countries that are eager for energy, and even individuals who are well-intentioned, and these people do not even know what they are doing, then this technology is easy. Abused or misused. This is why some Silicon Valley influencers and artificial intelligence researchers have come together to form OpenAI, an initiative that allows the world to use artificial intelligence, not just a group of special people. This measure is not "cheap" and it is reported that the quotation of a top artificial intelligence researcher in the project is two to three times its market value.
development of:
Another issue with artificial intelligence is the speed with which artificial intelligence can develop without being controlled by parameters. Universal Artificial Intelligence (AGI) has exponential growth ability because it will self-elevate and at some point will exceed the intelligence of its creators. The agricultural era lasted for thousands of years. The industrial era lasted for hundreds of years. The information age lasted for decades. Nowadays, the age of artificial intelligence only starts for several years. The growth of technology continues to accelerate rather than develop linearly, but now, any further acceleration It will make us unprepared to deal with the consequences (if we take a passive attitude, not a positive attitude). We have not yet been able to create super smarts, but we have mastered the processing capability - China's Tianhe 2 can calculate 34 cPs per second, far beyond the human brain (10 petaflops).
If we want to use this computing power responsibly, we need to have a foundation to deal with its unlimited development. This means understanding the ethics of artificial intelligence development, how to limit its use, and adopting appropriate protective measures when problems arise.
There are some serious concerns about how artificial intelligence is used and how it will develop in the future. However, it is irresponsible to excessively personalize the system being tested, or mislead us by the title. The invention of a new language by chatbots is not a threat. It is natural, and there are other more important things worthy of our attention.
(Selected from: venturebeat Author: TonyTie compile: NetEase see foreign intelligence platform compiler revision: Ginger sail)
China leading manufacturers and suppliers of DC Support Capacitors,DC Capacitor, and we are specialize in Electrolytic capacitor,High Voltage Capacitor, etc.DCMJ DC Support Capacitors
DCMJ DC Support Capacitors,Capacitors High Voltage,Dc-link Filter Capacitor,Dc Film Capacitors
YANGZHOU POSITIONING TECH CO., LTD. , https://www.cnpositioning.com