One-stop service platform of
High-end industrial weld pipe making machine

Home > News > Industry > Google AI Technology Using Norms: No Weaponization


You need a phone that we will provide the most suitable products, so you spend the least money to achieve the best results




Add.:No.13,Haiyu 2nd Road,Fuyu Village,Leliu Town,Shunde District, Foshan City,Guangdong



Google AI Technology Using Norms: No Weaponization

Source:Iris Liang Time:2018-6-1 10:04:03

Due to cooperation with the U.S. military on artificial intelligence projects, technology giant Google faced a surge of internal resignations and external public opinion pressure. Seko Machinery also pays attention to collecting daily processing datas. Based on these datas, heat preservation bright annealing machines could optimize its own operation.
In order to quell public pressure, Google promised to its employees that it will develop a set of principles to guide its artificial intelligence business in defense and intelligence contracts, and how to overcome ethical minefields. Google also said to the outside world that the new principles will prevent artificial intelligence from developing toward weaponization. However, it is not yet clear how to apply it in practical implementation.

Last April, the U.S. military proposed the Maven project. Maven's goal is to "accelerate the integration of big data and machine learning by the Department of Defense," and the US Department of Defense called the Algorithmic Warfare Cross-Functional Team (AWCFT). Previously, Google said that the project provided the US Department of Defense with a TensorFlow API interface for machine learning applications to help military analysts detect objects in images.

How did Google actually participate? Both parties did not mention it. According to the "New York Times" reported on May 30th, Google Cloud CEO Diane Green once spoke at the weekly company-wide TGIF (Thanks God, it's Friday) meeting. She said that the project was not for attack purposes. The contract value of the two parties is quite small, only 9 million US dollars.

Despite this, Google’s internal employees’ anger and disappointment have not been pacified. They believe that this cooperation has affected Google’s reputation and corporate culture, and it is also related to the moral issues of machine learning development and use.

As of May 14, local time, Google has more than a dozen employees formally resigned to Google for this project to protest the company's continued participation in the military's Maven project. At the same time, there are 364 artificial intelligence professionals in the academic community who also jointly sign and oppose Google.

Google was not prepared for the development of the situation. What makes Google even more troublesome is that the controversy caused by the Maven project may increase in the coming months.

The reason is that according to "Link" May 29th report, by 2018, the Maven project began to expand into new areas, including the use of development tools to drive hard drives more efficiently. In addition, the funding for this year's Maven project is also expected to expand to $131 million. The Pentagon also plans to establish a joint artificial intelligence center with Silicon Valley that may follow the Maven project to support the U.S. military and provide services to intelligence agencies.

On the one hand, they do not want to lose the military’s contract. On the other hand, employees question the company’s culture. Google faces a dilemma between “doing no evil” and “doing the right thing”.

In order to appease the emotions of internal employees, the "New York Times" said that Google held an internal seminar last Thursday (May 24th).

In fact, Google mentioned these principles and security measures at the beginning of the Maven project and it has not been disclosed to the outside world until now. Google employees expect the company to submit these principles and security measures in the coming weeks.

Faced with the company's internal resignation wave, Google originally had the opportunity to avoid serious situations when it came to project cooperation.

When Google discussed the matter internally in September last year, Google cloud chief scientist and vice president Li Feifei strongly suggested in the email that the two sides avoid the term "AI" (artificial intelligence) in the contract.
"No matter what price you pay, you must avoid mentioning or using the word AI. Weaponized artificial intelligence may be the most sensitive topic in the AI field. There is no one. For the media, this is simply red meat (red). They will find everything to destroy Google,” the scholar from the Stanford University artificial intelligence lab said in an email.

Finally, Google chose not to make the Maven project public. Only when opposing employees began protesting through Google’s internal communications platform, Google’s role as a subcontractor for the Maven project drew public attention.

At the beginning of April this year, opposition voices grew louder, resulting in the signing of more than 3,100 Google employees and submitting a letter to Google’s current CEO, Sundar Pichai, demanding that Pachachi cancel the participation immediately. The Maven project and drafting and publishing a clear policy show that: Google and its partners will not develop technology for war.

"Google's participation in the Maven project will irrevocably destroy Google's brand and its ability to attract talent. In increasing fears that artificial intelligence may be radicalized and weaponized, Google is gaining public trust. It has become apparent that the agreement will threaten Google’s reputation and will allow Google to stand on the opposite side of its core values,” Google’s employees wrote in the letter.

In addition to employee opposition, Google also encountered public pressure on the outside. On May 15, the International Robotic Weapons Control Commission (ICRAC) presented Larry Page, CEO of Google’s parent company Alphabet, Google’s current CEO, Sandal Pachachi, Google’s CEO, Diane Green, and Google Cloud’s chief scientist. President Li Feifei sent an open letter and once again demanded that Google stop its cooperation agreement with the Ministry of Defense on the Maven project, promise not to develop military technology, and guarantee that it will not use its collected personal data for military operations.

Controversy: How to define the use of technology

After a large number of employee resignations, Google also held discussions on the Maven project on April 11. According to employees who watched the discussion, Green firmly believes that the Maven project did not use AI for attack purposes, and Google AI researcher Meredith Whittaker found it difficult to define how to use the technology.

The primary task of the Maven project is to help the U.S. Department of Defense to effectively process massive amounts of video data from all over the world. The U.S. Department of Defense collects a large number of drone flight pictures every day, but human image analysts are far behind the increase in the number of videos. Therefore, the Department of Defense hopes to use machine learning to identify vehicles and other objects in drone footage, thereby reducing the burden on analysts. In the absence of its own resources, the Ministry of Defense has turned to academic and industry AI experts.

In this controversial project, TensorFlow provided by Google is a free open software in the field of artificial intelligence, and is also the key to Google's artificial intelligence strategy. It provides machine learning engineers with a framework for data sorting and training algorithms and is widely used throughout the industry.

But opposition voices said that new technologies are not always convenient, and researchers have warned that artificial intelligence technology may generate major biases in a subtle way.

Academics believe that Google has collected a lot of data on user behavior, activities and interests. Google cannot use this data alone to improve its own technology and expand its business. It should also benefit society. The motto of the company is "Do not be evil". This is a well-known responsibility.

Welcome to inquire us!
Iris Liang
SEKO Machinery & Technology Co., Ltd
E-mail: sales3@gdseko.com
Mobile phone: 13420628677
Tel: 0757-29390988
QQ: 845643527
Wechat/ Whatsapp: +86 13420628677
Skype: +86 13420628677

Related Product: S.S. Pipe mills,  High speed weld bead rolling machine, Industrial pipe making machine

[PREV]:How to Avoid Unnecessary Welded Pipe Scratch?
[NEXT]:Advantages of Welding Stainless Steel Pipes (2)

HOME | Tube Mills | Weld Bead Rolling Machine | Bright Annealing Machine | Cooperation case | Seko News | About Seko | Contact us

Add.:No.13,Haiyu 2nd Road,Fuyu Village,Leliu Town,Shunde District,Foshan City,Guangdong

Tel:0757-29390988  E-mail:seko@gdseko.com  Web:www.gdseko.com

Desigh by Ceall.net

Copyright © Guangdong Shunde Seko Machinery Co.Ltd. All rights reserved