AI Has Yet To Defeat Human Intelligence, Here’s Why: iFLYTEK CEO

2 min read

While the artificial intelligence industry has been around for 60 years, it wasn’t until AlphaGo’s landslide victory over Korean Go grandmaster Lee Se-Dol that AI returned to the spotlight.

Se-Dol’s match with AlphaGo brought more attention to the emerging sector, and it also put AI technology under tougher scrutiny from industry experts.

Some scientists argued that artificial intelligence might be the most apocalyptic technology of all. This view has an eager audience, including tech giants from Bill Gates to Elon Musk. The legendary astrophysicist Stephen Hawking went even further, remarking that “the development of full artificial intelligence could spell the end of the human race.”

Although Lee’s failure seems to support this ominous prediction, Hu Yu, the president of Chinese voice recognition giant iFLYTEK, believes AI technology, despite its quick development, still has a long way to go before it’s comparable to human intelligence.

iFlytech“AI is divided into three development stages: computational intelligence, sensory intelligence, and cognitive intelligence”, Hu Yu illustrated at a technology meetup held in Shanghai last week.

“Machines have outrun human brain in terms of computational capabilities, a point testified by AlphaGo’s victory over human Go master. Also, technology development in recent years has enabled machines to easily eclipse people’s sensory ability… Nobody can feel or detect the world through radar, ultrasound wave or laser, but computers can.”

“However, the most essential difference of human[s] from animals, or the factor that made us the ruler of the world, lies in our cognition, the ability to acquire knowledge, rationale and to understand through thought, experience and senses. No machine can do that now.”

“What’s more, all the AI-enabled machines are experts in a specific function, AlphaGo and IBM’s Deep Blue for board games, Pepper for companionship and entertainment. The list can go on and on, but none of the machines are made for general functions that are comparable to human being.”

As language is one of the largest distinguishing factors between humans and other livings, many scientists have used it to understand human cognition. “AI is out of the question without understanding of human language, the concepts behind the languages and the relations between the concepts,” he emphasized.

There are two mainstream channels to achieve revolution in sensory and cognition. One major and perhaps a little far-fetched way is to produce a computer replica of the human mind through cognitive neuroscience by simulating the organizing and working principles of the human brain. The tech industry has offered a more accessible solution to the same problem: a big data-based AI or artificial neural network.

The iFLYTEK Super Brain (讯飞超脑, our translation), an artificial intelligence project developed by the Chinese company, is among a rising wave of artificial neural network projects like Google Brain and Baidu Brain.

Currently, iFLYTEK Super Brain consists of a neural view machine, reading machine, and listening machine for sensory capabilities, said Hu. “Neural thinking machines for conception and decision-making and expression machines for execution are also integrated to achieve the leap from sensory to cognitive intelligence.”

Powered by iFLYTEK Super Brain, the company’s AI open platform now provides service to over 180,000 developers. According to the firm, AI technology is currently most favored by smart home and robotics companies, which accounts for more than 70 percent of the developers on the platform. Other industries ready to adopt the emerging tech include smart wearables, entertainment, and security.

Credit: 123RF Stock Photo