From 1958, when the American Atomic Energy Commission proposed plans to use nuclear explosions in buildings to Google Glasses, there have been many flash-in-the-pan projects and technologies in history. However, artificial intelligence (AI) is clearly not included, because it is constantly refreshing the sense of existence.
Recently, the British Journal Nature Outlook published a long article describing that AI is gradually infiltrating into our daily life, while people are fearing while enjoying it. In this mood, technicians who are vigorously promoting the next automation change need to face a serious issue: what the public really wants next.
Is there anyone else who is unaware of AI?
When the first mail begins to be sent digitally, the problem of sorting spam emerges. But the technology of identifying real mail from mass spam is also improving today. That's because smart spammers quickly found ways to avoid filters, so the e-mail platform started to use AI.
Now, after a lot of training, AI can complete the sorting work without human hands at all. At the beginning of this century, people worried that spam might strangle normal e-mail. Machine learning algorithm is very good at identifying the "routine" of mass email.
But we still need to look at the mailbox wastebasket from time to time, because machine learning is not perfect, and if we do not update the database in time, it is easy to be defeated by the spammer's new tricks. But in any case, things have become much easier. It can be said that machine learning is the best tool we have ever had.
The mail system is just one example. With machine learning, streaming media service platform can recommend movies to users, send goods that he may soon order online, and help identify people in photos or even flowers. Yes, these are AI.
AI drips in, but people are no longer unaware. Many people interact with computers every day, and Google's Alpha Go has impressed people by using machine learning to beat the human champion in a 3,000-year-old Go game.
Why are you afraid in the face of AI
Now, organizations and businesses are investing heavily in trying to apply machine learning to driverless cars. Obviously, this is a much bigger but riskier project than the simple pattern of identifying spam. So in this process, researchers have to face a problem: public perception of artificial intelligence.
This is actually a serious issue that every R&D worker who is vigorously promoting automation change needs to face directly - what the public is thinking.
In the field of artificial intelligence, science fiction has dominated people's ideas for the past century. The AI image in the novel has deeply influenced the public's view, so when facing its increasingly prominent position, people's general reaction is fear.
Part of this fear may be due to the fact that machines have perceptions that are not very different from human beings. In addition, the way AI research is reported may also cause panic. For example, in June 2017, AI researchers at Facebook said that two chat robots began using codewords in conversations. Some of the news reports even portrayed researchers rushing to stop the experiment in order to avoid things getting out of control.
In addition, warnings from prominent public figures may amplify fears. For example, Eron Musk, Iron Man, once said that AI might soon be strong enough to rule the world.
In fact, AI research in full swing has also had a cold winter. Around the end of the 1970s, there was a period of shrinking funding.
But in 1997, IBM's "Deep Blue" defeated the world chess champion and showed great computational power. Its purely violent solution shocked mankind and completely reversed the situation of AI research and development. This violent solution was later used in machine learning and easily won in seemingly innumerable games.
Now, the news that AI has been beating humans may give the impression that computers are comparable to humans in terms of cognitive ability. But in fact, there is still a gap between the two. Greg Holland, an expert in machine learning and natural language, tells people that the human brain can solve problems never seen before in AI, while machine learning is only designed for specific problems.
The Impact of Fear on Technology
Speaking back to the automobile industry, the rapid progress of artificial intelligence and the innovation of sensor system have brought dramatic changes to the industry. In the opinion of advocates of driverless cars, this technology can liberate people from the pressure of peak driving and reduce the occurrence of traffic accidents.
However, the industry's self-confidence contrasts sharply with the public's caution - technicians believe that driverless people are ready to go, but the public is not ready to ride. In April 2018, a Gallup poll showed that only 9% of adults in the United States would use a driverless car after the government regulators determined its safety; 38% said they would watch it for a while before launching it; and 52% said they would never want to use a driverless car.
The public's resistance to driverless driving is mostly due to safety concerns. Although statistics show that human errors are the cause of most fatal traffic accidents, robots are still not trusted.
In 2016, Tesla autopilot accident occurred. The survey found that drivers had overlooked repeated "hand steering wheel" reminders, but the public was still panicked. In addition, an unmanned car tested by Ultra Walk crashed into a pedestrian at night, allegedly when the AI on board did not promptly alert the car's security personnel.
Some manufacturers are trying to change that fear. GM's driverless cars will only travel in defined urban areas with clear roads and easy navigation, while in densely populated cities, cars will travel at half the speed on spacious roads. Other plans suggest that driverless cars will be restricted to the road when climate conditions are poor and AI's ability to navigate safely is hampered.
In any case, the biggest test remains the public's willingness to travel in cars driven by AI. As technology matures and rules improve, R&D personnel will soon see whether the public is willing to give AI this opportunity.
Source: Science and Technology Daily