本文章刊登在CHINA DAILY-艾问专栏



When Justine Cassell showed up with her first virtual human 18 years ago, what she met were only such interrogations.

In the following 18 years, she proved herself over time. “There are lots of virtual machines in the world and numerous lifelike figures in games. Can’t you tell the differences between the real and the virtual? Of course, you can. So please trust the human being and trust yourselves.”

At the “New Youth and New Future-2nd Conference on AI + Mobile Media”, Gloria Ai, founder of iAsk Media, brought us closer to Justine Cassell, Queen of AI.


Humans can Choose not to Move Towards the Singularity

Since the birth of Artificial Intelligence, it has been surrounded by such doubts and worries as what is the boundary between the virtual and the real, whether machines will become excessively powerful, and whether human will be exterminated by robots, etc.

No one is more concerned about AI than Google’s Chief Futurist Ray Kurzweil, who is acclaimed by Bill Gates as the futurist who could make the most accurate predictions about AI. He proposed the noted “Singularity Theory” a few years ago, warning about the impact AI may have on the human civilization in the future.

According to the “Singularity Theory”, the pure human civilization will terminate in 2045 when the powerful AI with infant-level intelligence comes into being. Then what is horrifying is that one hour after reaching singularity, AI will be able to deduce Albert Einstein’s Theory of Relativity and other fundamental theories that the human cognition is based on all at once. One and half hours later, this powerful AI will become a super AI with intelligence 170, 000 times that of an ordinary human. This is the singularity that will change the human race-the moment when computer intelligence and human intelligence merge.

However, Justine Cassell disagrees with the seemingly terrifying “Singularity Theory”. She believes nothing is inevitable and the key is whether we want machines to take over human. When we choose not to live in a world that machines take over from people, we will not move towards the singularity.


She believes that the human being will remain at the top of thinking capacity in the future as the manager and controller. The progress of AI will happen in the same way that numerous labors from the traditional agricultural and handicraft industry find new jobs in modern manufacturing and service industry. The most valuable AI technology is on how to combine AI with humans to make the human being more powerful and to form a symbiosis between human and machine.

Gloria Ai: Though the concept of AI has been raised for over 60 years, it does not become the focus of attention until in recent years. So what stage do you think AI is in now?

Justine Cassell: AI is currently very well known. Actually, it was popular 15 years ago. Artificial intelligence scholars often claim that AI undergo both summer and winter. AI summer is the time when progress is going on and everyone is excited. And then AI winter is coming. Few voices can be heard during then and nothing can be changed; people all get depressed. And then it comes back to summer, winter and all over again.

From Literary Lady to AI Queen

With brown curly hair and delicate golden-edge glasses, Cassell is so gentle and tender that many could not imagine her as an AI expert we commonly know. This also echoes to her definition as a “rule breaker”. “Althoughmy hating to complying with the rules has almost driven some people crazy, it is also probably the reason that I am able to manage the works that were impossible for others.The impossibilitiesin the eyes of others arethe great chances to challenge for me.”

That also explains the status that Cassell enjoys in the field of AI. As Associate Dean of the School of Computer Science at Carnegie Mellon University, she is reputed as the “Queen of AI”. Because of her outstanding contribution, she was appointed Chair of World Economic Forum (WEF) Global Agenda Council on Robotics & Smart Devices in 2011. And now she also serves as Co-Chair of WEF Global Future Council on Future of Computing.


Unlike most researchers in this field, Casselldid not flung herself into artificial intelligence at the very start. She majored in literature as an undergraduate, studied linguistics in her master, and finally acquired her PhD in psychology. It seems that what she has learned in her student days is not in any way connected to artificial intelligence. Her first “close contact” with computer science after several years working was simplyout of the expectation of completing the teaching tasks in a better way through the application of computer.

It was in the year she was invited to visit the University of Pennsylvania that Cassell truly came into contact with artificial intelligence for the first time. “I met a group of talented doctoral students and graduate students there. Together we created the very first virtual human. That is why I decided to go for interdisciplinary research.” Said Cassell.

It finally turns out that Cassell is indeed talented in computer languages. The Embodied Conversational Agent, her most prestigious scientific achievement of her as well as a pivotal achievement in human-computer interaction, deconstructs the language elements of human dialogue and confers human intelligence such as dialogue, narrating and social communication on machine languages. Having research of far-reaching influence in the industry, Cassell is now one of the authoritative representatives in the field of artificial intelligence.

Cassell has made many remarkable achievements in interdisciplinary research, but what makes heradmirable goes beyond that. As a female scientist, she has been encouraging women to participate in developing artificial intelligence. As a matter of fact, the lack of diversity and female participants are not uncommon in the global computer science and high-tech industries. The reportsrevealed that women only account for about 20% of the total number of engineersin large technology companies such as Google and Facebook.

Cassell told the iASK Media that creativity comes from differences and diversity, implying that the industry should involve not only male technical talents and artificial intelligence scholars, but also female professionals. The 50% of the students in her own laboratories are girls. She wants to make it clear to more female friends that you do not have to choose either being a scientist or a lady.


Gloria Ai: What makes you different from the rest of 7 billion people in the world ?

Justine Cassell: I am a rule breaker. That must be my most important trait. Although my hating to complying with the rules has almost driven someone crazy, it is also the reason why I am able to manage the works that were impossible for others.The impossibilities in the eyes of others arethe great challenges for me.

China Can Achieve a Great Success in Artificial Intelligence

“China is becoming more and more powerful in the digital world and artificial intelligence”.

Justine Cassell

China may not be considered as the main force of artificial intelligence 5 years ago. But we all realize that the development of artificial intelligence in China is very important to the world today.

The industry scale of China’s artificial intelligence was 6.93 billion yuan in 2015, and it reached 9.56 billion yuan in 2016.In accordance with the latest data from Tencent Research Institute, the industry scale of China’s artificial intelligence has exceeded 20 billion yuan in 2017.

Cassell is very confident in China’s artificial intelligence development. “I’ve been to Alibaba and talked with their leaders of artificial intelligence, as well as communicated with people in charge of Tencent and Baidu. They are doing great things. They know how to develop in these areas and how to train a group of gifted young people.As a result, they can truly make progresses in these fields.”


She also mentioned two major advantages of China in the field of artificial intelligence:

First, the most powerful parts of the current artificial intelligence technology are computer vision, natural language processing, recommended systemand network security. As one of the core technologies of artificial intelligence, natural language processing can never develop without Chinese language. Therefore, China’s artificial intelligence technology can be applied to large markets through products, algorithmsand data sets.

Second, the user data is private and difficult to obtainin the United States. However, China has a relatively relaxed legal environment in protecting citizens’ privacy, which provide a huge testing ground. The huge databases can be used to develop machine learning by some start-up companies while a large amount of data can support them in algorithm learning at the same time. As a result, Chinese companies can reach the levels that are unattainable for thecompaniesof same scale in United States.

The concept of “artificial intelligence” was born at Dartmouth College of the United States60 years ago. But China may have the opportunity to lead the development ofglobal artificial intelligence 60 years after the formation of concept.

Gloria Ai: More and more capitals and companies are now doing everything to connect themselves with the concept of artificial intelligence. This may be a positive power for the development of AI, but does this phenomenon also mean the existence of bubbles?

JustineCassell: We must remember that artificial intelligence is not a technology, but a paradigm and a method of manufacturing machines that can perform all human behavior. If you simply add AI to the front of a company name, it doesn’t make any sense and means neither to put machine learning in practice nor to analyze the customer’s data. In short, AI is like an umbrella, below which are many running components that you cannot see.