Cutting edge technology

Whether you are ready or not, a large wave of artificial int

Introduction: Artificial intelligence has been around for decades, and in the public's imagination, it has become one of the iconic technologies of 2018, although it does not always show people in a positive way.

Artificial intelligence has been around for decades, and in the public's imagination, it has become one of the iconic technologies of 2018, although it does not always show people in a positive way.

On the plus side, artificial intelligence and related technologies (such as machine learning and deep learning) can be used for speech recognition on smartphones and devices such as Amazon's Echo and Google's Home. Services have become a matter of course, autonomous vehicles, more accurate disease diagnosis, and less obvious but at least equally influential, more automated information technology infrastructure in the cloud and data centers.

At the same time, artificial intelligence is used to create fake news to attack people, discriminate against certain types of workers or customers, and cause fear, although this may be too exaggerated, and the machine may soon eliminate most jobs. Tesla’s CEO, Elon Musk, and the late physicist Stephen Hawking, among other important figures, have caused particular concern (people are still arguing over this,) that out-of-control artificial intelligence may threaten To the survival of mankind.

But whether the outcome is good or bad, the new year will undoubtedly accelerate the use of artificial intelligence and machine learning in a variety of products, businesses and daily activities. The following are some predictions about what is going to happen and what will not happen, and the ideas put forward by experts:

The importance of artificial intelligence in the enterprise field is no less than that in the field of consumer products

[123 James Kobielus (who is the lead analyst for artificial intelligence, data, data science, deep learning, and application development) at Wikibon, a similar market research firm in Silicon Valley, says artificial intelligence is reshaping business intelligence. This allows business users to perform a large number of analyses that were once trained by trained data scientists.

There is also robotic process automation (or software that simulates how people perform tasks in the process), which has become one of the main enterprise use cases for artificial intelligence. Artificial intelligence is also becoming an important foundation for managing information technology infrastructure. This is called "Artificial Intelligence Operation and Maintenance (AIOp).The emerging model of s). As Kobielus points out, the idea is to make infrastructure and operations more capable of self-healing, self-managing, self-protection and self-optimization.

In particular, machine learning Gradually, software development itself has changed in such a way that machines (rather than developers who have to write specific logic and rules) can create applications. This will become more apparent in 2019, with cloud computing. This is especially true for giants providing more and more artificial intelligence services. In the recent Dun & Bradstreet survey, nearly half of companies said they are deploying artificial intelligence systems, and another 23% are planning. Stage.

What other people think

"In 2019, more business intelligence (BI) suppliers will integrate a large amount of artificial intelligence to automatically extract prediction insights from complex data. And provide these complex features in the solution that can simplify self-service and be the next best action (next -best-action) Provide guidance for countermeasures. "When Wikibon's James Kobielus said.

"Machine learning will enter the operational phase, from behind the scenes to the front of the stage, into the real-time and critical enterprise application structure. ”Monte Zweben, CEO of Splice Machine, said.

“Don’t tell me about one or two artificial intelligence projects you are doing. I think hundreds of projects. "" IBM's analysis general manager Rob Thomas said at theCUBE summit.

"Humans in the loop" will become the norm - but not always realistic

[123 Because Amazon's artificial intelligence-driven services such as Alexa tend to perform well, it is believed that artificial intelligence will take over various jobs. The truth is far from this, of course, it will certainly become a reality soon. McKinsey estimates that there are less than 5% of jobs. Automation can be fully implemented using existing technology, but at least 30% of activities in approximately 60% of occupations can be automated.

All this means,In the years after 2019 or 2019, the most successful applications will be those that help people get the job done better, whether it's a clinician who analyzes through MRI scans or a factory worker who tries to handle more potential customers. , industrial robots or mortgage lenders.

That is to say, one's high yield is often at the expense of another person's contribution. In view of this, some people stubbornly believe that artificial intelligence is only a tool, and this view seems a bit hollow. If artificial intelligence really needs to benefit society without exposing a large number of people to unemployment, then artificial intelligence providers and companies using artificial intelligence must prove this in 2019. Private companies and governments must come up with solutions as soon as possible to help people who are unemployed because of efficient artificial intelligence.

What other people think

"In 2019, artificial intelligence will continue to make our work life easier and allow us to do more things... workers will be responsible Some tasks or projects are handed over to the machine according to our preferences." - David Leonge, vice president of SAP for SAP Leonardo, machine learning and intelligent process automation, said.

Artificial intelligence will become more transparent as faults and fears intensify

Machine learning has suffered a major blow (especially the use of artificial neural networks for deep learning and other types of things The blow), that is, the algorithm used to produce the result is a black box. That is, you input a lot of data, but the results you get are not always clear - sometimes even wrong, such as when an autonomous car unexpectedly stops on the road because of an inconspicuous object, but sometimes But failed to correctly identify humans and kill people.

In fact, in a recent survey conducted by D&B, almost half of the respondents said that the interpretability of artificial intelligence has caused problems for the organization, with 46% of respondents saying that At least they have trouble in this area, figuring out how the artificial intelligence system gives the answer.

Equally bad is that the data used to train artificial intelligence systems is wrong or biased. For example, in 2015, Amazon had to give up using the tool when it learned that its artificial intelligence-driven recruitment tool was patriarchal because the tool considered most of the applicants employed.The number is male, which means that men are superior. This year, this understanding may translate into more action, thereby avoiding the occurrence of such things – resorting to legislative means if necessary.

Although we can hardly spy on the shackles in the black box, just as we can't understand people's brains to analyze the decisions they have to make, people's need to understand the internal working mechanism of artificial intelligence is becoming more and more urgent, especially legal work. By.

There are no doubt that some of the technology companies that have solved this problem as a unique advantage will not lead the trend. The government may ask the company to implement some degree of transparency, but it is not clear how the government can do this. But this will be a bigger issue this year.

What other people think

"The global technology giants who have mastered artificial intelligence and have great energy have raised a lot of questions about how to regulate the industry and technology. In 2019, we must think about these issues. The answer - since technology is a multi-functional tool that brings different results with the specific use environment, how do you standardize the technology? How do you formulate it will not stifle innovation, or do not favor large companies (large companies can afford Compliance costs, and small companies can't afford the regulations? At what level do we regulate? International, national, or local level?" - Managing Director and Responsible AI initiative of Accenture's Application Intelligence Division Rumman Chowdhury, global head of the company, said.

"At the time of the accident, the division of responsibility may have to be resolved in the court. New legislation must be enacted so that the court can obtain sufficient reference on the thorny issues of responsibility." - Chief Analysis of Mobilocity LLC Teacher J. Gerry Purdy says

"Maybe we should learn from human psychology" to make artificial intelligence easier to interpret - Danny Lang, vice president of artificial intelligence and machine learning at Unity Technologies.

"2019 will be a year of action. People will make more commitments and statements, namely the responsible creation and use of artificial intelligence, and the company will have to adopt it. Influencing human rights In the decision-making process, the public willOppose the use of artificial intelligence that would create bias. More employees will demand an impact on what they have created and refuse to contribute to harmful automation. Companies must set their example with conscience (whether buying or creating artificial intelligence solutions) and try to ensure that the system is fair and equitable to avoid becoming the next headline of AI problems. "Kathy Baxter, Architect of AI's ethical practice at, said.

"Congress will slowly oversee AI, which requires more certification at both ends of the consumer and business." Information, country of origin information and transparency are just a matter of time. In particular, banks must be wary of discriminatory practices associated with the use of big data, and banks must constantly assess the prejudice that may be brought into the algorithm by those involved in the development. "--Kayvan Alikhani, co-founder and CEO of"