Shulou(Shulou.com)06/02 Report--

As ordinary people, most of our intersection with AI lies in the mobile apps of mobile phones. Jinri Toutiao's information stream recommendations, mobile phone photography of AI photos, including prisma, the main neural network migration of the "hard-core AI app." Recently, however, more and more AI applications have begun to enter the PC side. One of the major positions is office software.

It can be seen that more and more office software begin to apply AI technology. For example, the Adobe series has applied technologies such as video intelligent tagging classification, intelligent editing and automatic content recognition, while Office 365will automatically use the data in Excel to generate charts, and automatically sort, sort and summarize emails.

But what we are going to talk about today is not the office software AI as a productivity tool, but the AI tools that are entering the workplace communication and exist as a "workplace lubricant".

E-mail, recruitment, collaborative work, AI how to tear the palace fight into laughter?

The word "communication" has always been a very special existence in the workplace. In a small way, the way of communication is only some optional "e-mail etiquette", but many times the problems in communication often become the fuse. For example, an email reply is not polite enough, which is likely to affect the customer's dissatisfaction with the entire project team.

Even in the window of external communication, such as recruitment announcements and public notices, once problems arise, they will brew into a public relations disaster. For example, when an Internet company mentioned geographical issues in its recruitment message in 2017, it was widely criticized for directly "going out of the circle". Whenever an enterprise tries to conduct public relations by means of open letters and public apologies, it must be followed by a team of "fault-picking sentences".

In short, the "devil theory of details" in the workplace is often hidden in those casual language. After all, people, as perceptual animals, work is a relatively rational and closed environment, so emotions can easily be revealed in the relatively gray link of communication.

Of course, AI is the only employee who wants to say "moody". We can look at several cases to see that AI optimizes the communication process.

First of all, in terms of e-mail, Google's Gmail is a typical case.

In 2016, Google Gmail launched Smart Replies-- intelligent reply function, which uses recurrent neural network to structure large amounts of email content data, learn users' language habits, and provide some automatic reply suggestions according to the context of the email.

Last year, Gmail added Smart Compose, smart email writing function to the smart reply. Automatically give writing suggestions when users write emails, and these suggestions are very polite "polite words". For example, AI can extract the recipient's name through the mailbox name and automatically add "Hey" at the beginning; it can also guide the user to blow rainbow fart to the recipient according to the content of the email, for example, if the user hits "Thank you …" AI will guide users to type "it really helps me a lot"; if the delivery time is Friday, AI will also advise you to add "have a good weekend" at the end of your email.

Microsoft, on the other hand, has come up with a more exaggerated idea that in the future AI may identify negative emotions in users' emails and ask again before users press the send button.

This intelligent email etiquette can bring many benefits.

First of all, it at least saves the time for users to enter these polite words, and at the same time, it can also serve as a buffer and guide when users do produce negative emotions. If users are newcomers to the workplace, this mechanism of automatic data training from personal mailboxes can also help them adapt to a corporate email etiquette style more quickly.

In addition to emails that focus on a small range of internal communication, some companies are now starting to use AI to solve external communication problems such as recruitment advertisements.

Like Textio, which has always focused on the enhancement of business writing, the original product is the AI optimization of business copywriting, but in recent years it has begun to focus on using the same technology to serve corporate recruitment. Users can submit their own recruitment copy to Textio, which is rated by the Textio system and is generally released with a score of more than 90 points.

Through the training of recruitment copywriting, especially the job description and recruitment feedback data, Textio can also provide users with a set of career description writing guidelines. What kind of position should be accurately described and which words will increase the pressure on the applicant can be instructed and corrected by AI.

The effect of using Textio is obvious. Customer Zillow Group says the company has improved its efficiency in processing recruitment messages by 16%, while Johnson's data has increased by 25%.

Even the enterprise communication and cooperation platform slack, which is used to integrate various kinds of chat, email, text messages and groups, has begun to appear slackbot to help users deal with messages. Developers have also added a large number of natural language processing technologies to the development of slackbot. For example, many people will distribute work tasks to colleagues who collaborate remotely through slack. With the introduction of slackbot, more modal particles, politeness and other words can be integrated into the task distribution, which makes the whole task distribution process more humanized.

Generally speaking, when more and more machine learning and NLP technology are used in workplace communication, it can really save a lot of time to think about words and avoid some unnecessary friction to some extent.

Where is the boundary of plastic AI love?

But in fact, not everyone welcomes this "workplace lubricant".

The most important problem is that AI is quietly expanding the boundaries of its applications in this way, gradually spreading to emotions from a physical point of view.

In the case of emailing, for example, even if it is inappropriate to show your negative emotions in an email from a workplace point of view, it doesn't mean that tech giants have the right to guide or even control our emotional expressions.

If the tech giants have calculated that cheerful and lively language is more conducive to productivity, will they also use enhanced writing technology to guide employees to write down seemingly happy emails against their will?

In the long run, will the true emotions of employees be suppressed, thus triggering more negative reactions?

By learning the language habits of different groups, AI has been able to "watch people eat dishes" and know how to use language to arouse each other's emotions. Take Textio, for example, one of the services Textio can offer is to analyze which words in recruitment copywriting are more likely to attract male job seekers, such as "ninja", and which words are more likely to attract female job seekers, such as "independent". Many people believe that the advent of this technology is making workplace discrimination invisible, and if a company does not want to hire women, AI can help write a "straight male" career description that makes women uninterested.

What is even more frightening is whether this technology will move from the workplace to more places when it comes to emotional manipulation and simulation.

Google Gmail's smart reply model may launch more models to help users simulate a tone aimed at couples, relatives and friends. Textio's enhanced writing may move from business to daily life, helping users write New year's greetings, love letters and apology letters between relatives and loved ones.

On that day, is it true or false that our words and words express?

We often say that AI can help human beings to complete repetitive labor work, liberate human energy and time, so that we can better devote ourselves to creative work.

But when one day, AI can replace us to output emotion through language, will human beings become more affectionate? Or is it more ruthless?

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.