"We have all been exposed to or learned about scenes of virus release. The vitor's performance process is like the self-propagation of the virus..."
"It's not even as good as the self-diffusion of the virus, because the vitor can't do anything that exceeds the budget, and the virus may fission into mutations..."
"So, there is no loss of control and no intelligence. The program just releases what it has learned. It is still far from the threshold of strong intelligence."
Su Xiaomu concluded.
However, Su Xiaomu also knew very well that as it was now, he would definitely not be able to convince these top players in the world.
The next step is undoubtedly to answer questions.
Explain intelligence, vitor, etc. in more acceptable terms.
The first person to ask questions was Professor Li Feifei. She has always had some meaningful ideas about artificial intelligence.
Including the co-director of the Stanford Institute for Human-Centered Artificial Intelligence, which she currently serves as, she also led and initiated.
"A brand new angle, but I have a doubt. Judging from our current observations, vitor's performance already contains a little bit of diversity in human wisdom. I can understand what you said about releasing learning, but from what angle?
Do you think there is no self-fission inside?"
"Perhaps the Turing test can give the answer." Su Xiaomu smiled and said relaxedly.
The Turing test is still the most reliable testing standard currently available.
At least, Turing's prediction in 1950 about the thinking ability of machines in the year 2000 has not been achieved until now.
Su Xiaomu's proposal was recognized by everyone.
Li Feifei looked at Su Xiaomu curiously, which meant a lot.
Since there are so many outstanding researchers, this Turing test is very strict, with many test scenarios and testers arranged.
According to Turing's prediction, if more than 30% of testers cannot determine whether the person being tested is a human or a machine, then the machine will pass the test and be considered to have human intelligence.
Fortunately, there are enough people in the experimental group and their academic attitude is firm enough.
No fool can join such an experimental group.
Just like a concept expressed by Li Feifei himself: Academia is inclusive and does not compete for market share. It requires people to work together to obtain a deeper understanding of knowledge, and this knowledge can be shared.
What's more, academia is responsible for educating future leaders and practitioners, across a range of disciplines.
The development of artificial intelligence will be a multi-generational journey, and now is the time to instill human-centered values in technicians, engineers, entrepreneurs, and policymakers, because it is these people who are responsible for charting the future course of artificial functions.
The first and most important point is that the academic community is inclusive and does not compete for market share, so in the current scenario, it is difficult for someone to deliberately express wrong opinions.
Even if he himself is willing to sacrifice all future academic possibilities, everyone will not recognize the validity of the test results.
"..."
There is nothing to say about the testing process.
These are some classic Turing test questions.
As the test progressed, the test results became increasingly disappointing.
All in all, there are very complicated emotions here, including Professor Li Feifei, Professor Grosz and others. They all actually hope that strong intelligence will emerge.
But I don’t want it to appear so suddenly that there is no accurate and objective judgment.
It’s as if Li Feifei’s human-centered artificial intelligence has outlined three simple yet powerful ideas:
1. In order for AI to better serve our needs, it must embrace the diversity, nuance, and depth of human intelligence.
2. The development of artificial intelligence should be combined with long-term research on its impact on human society and guided accordingly.
3. The ultimate goal of artificial intelligence should be to enhance our humanity, not weaken or replace it.
And the question that Li Feifei himself has mentioned countless times: How can we make intelligent software consistent with human value orientation?
Before getting definite answers to these questions, they didn't want anything unexpected to happen.
In fact, just when everyone was thinking about similar questions except for the depth, they already had the answer in their minds.
Vitor is definitely not a strong intelligence.
At least vitors are not as complicated as humans to have a million thoughts in one moment.
As the test ended, everyone's interest became less and less. It was almost noon, and MIT had already made arrangements to receive the results.
Whether they were members of the experimental group or the observation group, they all left the experimental group in groups under the guidance of the corresponding personnel.
Su Xiaomu, who was supposed to go with everyone, was not inside, but was left behind, and then followed Tommaso to the office.
Also accompanying us were Professor Grosz and Professor Li Feifei.
There is no one else.
"Old friend, you actually don't understand either, right?"
As soon as the office door closed, Professor Grosz looked at Thomas Ropeway with a smile.
Li Feifei showed the same look, and apologized solemnly: "Professor Tommaso, I'm very sorry. I was so excited that I lost my temper... I deeply apologize for the impact on you..."
Tommaso smiled and waved his hand: "Don't worry about some small things."
"In fact, I didn't expect that there would be so many surprises just from yesterday to today. Another model of realizing intelligence created by Su has great potential!"
"Sure enough!" Grosz said with a smile.
"Su, stop standing and sit down. We three old guys all want to hear your redefinition of intelligence." Tommaso looked at Su Xiaomu.
Grosz and Li Feifei also looked at Su Xiaomu.
As early as Su Xiaomu saw the vitor's unexpected performance and started to explain it himself, he had expected that a more detailed explanation was needed.
"Okay, let me tell you my personal views. I just want to represent my personal views on intelligence, which may be different from the popular definition."
Su Xiaomu took a precaution first, and Su Xiaomu was already very good at using the knowledge gained from the little trash for his own use.
"Different from the previous understanding and implementation of intelligence, as I just said, in my opinion, intelligence has actually appeared a long time ago. The spread of viruses and independent diffusion of fission is a crude manifestation of intelligence. Note, I
It’s talking about strong intelligence.”
"We all know that human thinking is diverse, changeable, multiple, subtle, difficult to predict, and unable to maintain absolute sanity. However, no matter what kind of data it is trained on, the intelligence will faithfully perform the set operations, some subtle
Expression can be mistaken for an advanced form of intelligence.
Actually no.
Using algorithms to dominate programming and intelligence will increasingly lead to strong intelligence, because we do not train programs, but let them develop linearly..." -
