Is AI going down?

Original article was published on Artificial Intelligence on Medium


After Google’s Deep Mind AlphaGo beat South Korean Go champion Lee Se-dol in 2016, AI grew to become a countrywide subject and entered human productive life.

In 2015, capitalists have begun to compete for AI. According to facts from the Investment Research Institute, the scale of financing in China’s AI enterprise in 2015 was once 45.827 billion yuan, with 489 financing events, a considerable enlarge from the preceding year; the scale of the financing in 2016 used to be 48.926 billion yuan with 607 financing events; and in 2017, it rose to 72,726 million yuan.

Bob (Xiaoping) Xu, founder of ZhenFund, a China-based seed stage targeted project fund, even said: if it used to be Internet+ a few years ago, then the subsequent 5 years have to be AI+. For this reason, he believes that “the cash invested in AI is now not enough”.

There had been 3,942 newly registered agencies in China in 2015 involving the key-word “artificial intelligence,” and the range of new registrations reached 5,070 in 2016, an amplify of 28.6%, in accordance to Qichacha.

Among them, the greatest quantity of these are pc imaginative and prescient (CV) companies. And a range of AI unicorns had been born, such as SenseTime, Megvii, CloudWalk Technology, YITU Technology.

In 2017, AI functions have blossomed from laptop imaginative and prescient to deep learning: Intelligent Connected Vehicle (ICV), clever carrier robots, clever speakers, and clever domestic merchandise have been created. Soon after, a number of Internet giants additionally started to pile in, and the enterprise grew to become extra competitive.

First, in June 2017 Tencent opened its AI applied sciences such as laptop vision, wise speech consciousness and herbal language processing to the public, which unexpectedly reduced the entry threshold of the industry; Baidu then launched two open platforms, DuerOS and Apollo, beneath the banner of “All in AI”, with the intention of occupying the excessive floor of AI running gadget and driverless; afterwards, Alibaba additionally joined in and introduced the institution of DAMO Academy, investing closely in trendy applied sciences such as quantum computing and computer learning.

AI entered its heyday in 2018, pushed with the aid of the giants.

In September, the 2018 World Artificial Intelligence Conference was once held in Shanghai and President Xi Jinping for my part despatched a letter of congratulations.

This conference, with the mission of contributing “world solutions” for AI, is a focused show of the achievements of China’s two years of fast improvement of the AI industry. AI+education, AI+finance, AI+manufacturing, AI+autonomous driving, and AI+health are the seven “AI+” themed experiences, which for the first time comprehensively exhibit how AI will be utilized to life. Since then, AI has begun to emerge as a new label for China’s technological innovation.

The accelerating arrival of the AI technology is surely heartening. But under the excitement, crises and dilemmas loom.

One of the most frequent examples at that time used to be that entrepreneurs ought to take a PPT of any AI-related assignment and then appear for traders to get funding. The task that has obtained financing is continuously speculating new ideas for excessive valuation, and touchdown as an alternative will become secondary. Eventually,valueless purposes and overheated capital have led the fast-growing AI to a bubble.

In early 2018, Kai-Fu Lee, AI Expert, CEO of Sinovation Ventures, envisioned on social media that the AI Bubble (2018) ) to destroy by way of the quit of the year, AI wintry weather was once coming.

Sure enough, at the cease of the year, Kai-Fu Lee’s prediction got here authentic — the AI iciness used to be here. The first is that funding and financing in the AI discipline turns into difficult. The analysis of lookup company ITJuzi suggests that the AI enterprise raised $85.4 billion in funding in 2019, down greater than $30 billion from the preceding year; the quantity of financings additionally fell to 527 from greater than 800 the preceding year. Cautious capital flocked to the head companies, different boom stage AI organizations are extra tough to survive. And even some traders are beginning to flee the AI track. Right after that, some AI groups began to stagnate or collapse.

In the subject of pc vision, which used to be as soon as the most sought after by means of investors, even the head agencies are additionally trapped in commercialization problems, whilst the survival of smaller organizations is extra difficult. In the driverless field, Roadstar.ai, a driverless startup valued at $400 million at the starting of 2019, has been a important success due to the team’s infighting swiftly closes down, kicking off a iciness in the driverless industry. Since then Drive.ai has been acquired, Baidu Apollo is making sluggish progress… and driverless discipline is beginning to go downhill.

According to Qichacha, the range of organizations with key phrases containing “artificial intelligence” written off/suspended in 2019 is 2565, which is 5 instances greater than the preceding year.

Capital and Genius are poured into the AI field, so why is there such a situation? An AI entrepreneur as soon as wrote an article revealing the motives for the commercialization difficulties of AI companies: first is the sturdy science and susceptible product sales, which are some distance from the actual wants of users; second, the algorithm is susceptible to upstream and downstream pressures, which can without problems lead to a waste of money; third, AI science is in truth a To B area and its enterprise increase is slow.

As for the driverless field, in addition to technical bottlenecks and group conflicts, the market is no longer ready, even worse, more than one driverless accidents make human beings concerned about it. And the separation of autopilot corporations from OEMs has additionally created a large gulf between R&D and mass production.

On a broader societal level, AI is additionally plagued with the aid of protection and moral issues. On the one hand, driverless deaths, AI fraud, AI privateness leaks and different incidents have took place one after another, AI science at the back of the “black box” constantly make humans afraid of it; on the other hand, AI, which hasn’t made any cash yet, has been concerned in a couple of political and social occasions such as the U.S. election manipulation, Facebook spying, racial discrimination in Amazon’s hiring algorithm, and Google’s improvement of AI technological know-how for army strikes.

In distinction to the second of scenery, AI faces a sequence of problems, such as the disappearance of technological dividends, limitations to business landing, and protection moral issues.

The AI draw back grew to become the new image of the industry.