Sina Technology Zheng Jun from the United States Silicon Valley
This year'sGoogle I/O ConferenceThere is no suspense, and suspense. The reason why there is no suspense is that the only subject of the conference is definitely artificial intelligence. Nearly all new products are run through machine learning technology. The reason for the suspense is because the outside world is expecting and paying attention to this technology giant. How new heights will AI technology be pushed to.
Behind human nature is privacy
The many new software products released yesterday have opened up new horizons in the AI era. Gmail can automatically predict and help users compose emails; Photos can automatically help users sort and retouch maps; Assistant can not only understand the user's two different instructions at the same time, but also can automatically help users to book restaurants; News can automatically build news topics for users. Providing the ins and outs of hot events; Lens can automatically recommend and convert texts for users. And so on, the list goes on.
Of course, the coolest scene in the audience is the automatic booking of assistants. The user only needs to tell Assistant that the AI will automatically call the merchant to book in the background. In the live demonstration, the audience was basically unable to hear the call and was actually an AI robot. The natural fluency of the dialogue can almost be real.
This amazing feature comes fromGoogleThe Duplex Voice Project will be launched later this year. Although the restaurant's haircut was displayed on site in a seamless manner, Google did not blow up the ceiling. Duplex is pre-trained specifically for specific scenarios and is currently only used for specific purposes. Understandably, this robot will also be guilty of dizziness if it encounters a special, persistent business. Google intends to use this feature for a variety of telephone booking services to get more data for training.
The most needed of machine learning is data and training; and massive users and data are the biggest advantages and foundations for Google to promote AI research and development. Google has a billion-user base product: Android global users surpassed 2 billion last year (excluding mainland China), and many products such as search, email, video, browsers, and maps exceeded the 1 billion level a few years ago. Google provides a complete set of Internet services for users around the world at no cost, and it also acquires massive amounts of user data for its own advertising services and AI R&D.
On the other hand, users who want to enjoy the many human and convenient experiences brought by Google's AI technology will need to increasingly entrust their own personal data and online behaviors to Google. In other words, how much do you believe Google can enjoy in your convenience? In the era of AI, Google is becoming the user's webmaster, gaining more and more user data and privacy.
Specific examples. Gmail automatically composes emails for you. It means that Google knows your address book, address, and telephone information in your address book. It collects your usual habits of speaking and wording. Photos automatically categorizes and edits photos for you, meaning that Google knows your photos. All the friends in the room, Assistants are considerate and caring, meaning that Google knows your spending habits, travel history, and payment information; News knows your preferences and concerns; Lens knows what objects you took. Of course, there are Autofill features that hold all your passwords. In a nutshell, Google knows almost all of your secrets on the Internet and on your mobile phone, if you authorize it.
Li Yanhong's confidence and embarrassment
Interestingly, I remember two months ago,BaiduLi Yanhong, founder and CEO, publicly stated that "Chinese users are willing to give up privacy for convenience." This has caused a lot of public anger and even official media criticism. His original remark is that "Chinese people are more open to privacy issues and relatively less sensitive." If they can use privacy for convenience, security or efficiency, they are willing to do so in many cases. ”
In fact, there are still two words behind Robin Li, and we must not take it out of context. "Of course, we have to follow some principles. If this data can benefit users, and they are willing to use it for us, we will use it." I think this is the basic criterion for what we can and cannot do. ”
Simply speaking about the matter, there is no problem with the full meaning of Li Yanhong’s speech. This was originally a positive cycle of an Internet company promoting human services and AI technology. Users trust Internet companies and authorize their own data. On the premise of ensuring the security of user data, Internet companies provide convenient services based on data, use data for machine learning training, and use data to obtain advertising revenue. This is exactly what Google has done. It is Google's business model and is also the driving force behind Google's AI technology.
So, why did Li Yanhong say that this statement would be irritated? The reason for this is obviously everyone knows. Privacy has always been a sensitive topic. Even if the business model is such, Google does not necessarily dare to say this, let alone Baidu. Previous series of negative incidents that excessively pursued commercial interests while ignoring public responsibilities, especially the issue of false medical advertisements that had repeatedly been banned, had severely damaged Baidu’s brand image and public trust.
Li Yanhong possesses knowledge, technology, career, and value. He is also the most successful Silicon Valley returnee entrepreneur. If it is not Baidu's endless variety of negative aspects, he should be the most high-quality and sunny entrepreneur idol in China's Internet industry. Similarly, Baidu is also the leading Internet giant in technology research and development and storage in China. It is at the forefront of the Chinese Internet industry in cutting-edge technology such as artificial intelligence and automated driving. At the Baidu AI Developer Conference in July last year, Li Yanhong sat on the Fifth Ring Road in Beijing with an autonomous vehicle. The scene of laughter was impressive.
But never-ending negative events have made Baidu’s brand image shy again and again. The Wei Zexi incident is the lowest point of Baidu’s public image so far. When Li Yanhong was interviewed by Caijing Magazine in June 2016, he said that he would also go online and read the comments of the public on Baidu. In October of that year, when he was speaking at the China University of Science and Technology, he was also met with protests by students who had not been angry. Li Yanhong had nothing on the stage, Baidu search also showed nothing happened.
I believe that Li Yanhong is not a corporate executive who doesn’t hear anything from the window. It only shows that he believes that the public has forgotten the negative aspects of Baidu and has strong self-confidence in the image of himself and Baidu, and will make embarrassing remarks one after another. . As he publicly stated, Google did not withdraw from the Chinese market because of competition. In fact, Google’s search market in mainland China once had a 30% market share before Google’s high-level impulses.
The AI Times Should Be More Responsible and Ethical
A consumer-oriented Internet company whose business is based on users and data. Getting the user's data and trust is a heavy responsibility behind it. An Internet company dedicated to artificial intelligence needs more responsibility and ethics. It must first be a company that values safety, ethics, and responsibility to ensure that AI technology is used in good faith. Otherwise, the abuse of technology and the leakage of data will bring huge losses to users.
Zuckerberg at the hearing site, data map.
In the face of adverse user privacy protection, social giantsFacebookHas paid a heavy price. Third-party developers privately sold user data, putting Facebook under heavy pressure from regulators and once evaporating $50 billion in market value. Zuckerberg had to publicly apologize to the public several times and personally went to Congress to receive a ten-hour wheel-wheeling hearing. In contrast, I did not see Li Yanhong having the same apology sincerity.
Google CEO Picay, data chart.
inMicrosoft Build ConferenceAt the same time as the Google I/O conference, the two big technology giants emphasized the importance of protecting data security and user privacy while talking about artificial intelligence.MicrosoftCEO Nadella spent a lot of time talking about the three principles behind AI——privacy, responsibility and ethics, and formed an AI ethics committee at Microsoft. Google CEO Picachai said that technology can be a positive force, but (Google) feel deeply responsible for the correct handling and application (technology). This year's Google I/O Conference talked about the use of many AIs in the medical and health field. The new system, Android P, has spent a lot of effort on improving digital health and reducing the user's indulging and relying on smart phones.
In fact, holding high "Do not do evil", although the banner of Google does not have such a negative burden of Baidu, but each of their technical updates, are faced with scrutiny by the regulatory authorities. It can be said that Google’s not doing evil is more from the head of the “Sword of Damocles” hanging over its head. It comes from repeated punishments; it is the government’s regulatory pressure that has caused Google to dare not. Deliberately for user data.
In 2011, Google was penalized by the U.S. government for US$500 million for illegal prescription drug advertisements. In 2011, Google agreed to accept the US Federal Trade Commission (FTC) for 20 years because of its illegal promotion of its Buzz social product; The Safari browser violated the rules and obtained user privacy and was subject to a penalty of US$22.5 million. In 2013, Google Street View’s collection of user data was penalized by the US’s US$7 million. The last time it cracked down on the European Union, which had a culture of robbing the rich and the poor, Google had been punished by the EU's 2.42 billion euros for the price of its own web pages in search results.
The vast majority of AI products released at this Google I/O conference cannot be used at home. The domestic internet giants are also developing their own AI technology. In terms of convenience of life, the domestic giants are even better than Google. For example, Gaode Baidu Maps is superior to Google Maps in terms of living services. However, the first thing behind artificial intelligence is responsibility and ethics. It is hoped that the domestic internet giants will also be able to assume the responsibilities and ethics of user data while learning advanced technologies in Silicon Valley.