The brain of the company. Part 3

The continuation of the story about the ups and downs of the implementation of AI in a trading company, about whether it is possible to completely do without managers. And what (hypothetically) this can lead to.



Bots decide everything



- Max, I congratulate you, we have done almost everything on the sales chain. There are improvements and receive a percentage for three years, as prescribed in the contract.

- This is only half the project. We have not yet taken up the most important thing.

- Wait, what's the main thing? What for? We have done everything!

- We have automated processes in the sales chain, everything works well without people, but there are no more customers. They must be attracted to their side on the Internet. We need to make bots.

“But we did the perfect service, the customers will appreciate it and come by themselves.”

“They don't seem to be in a hurry, but I have no time to wait.” Not interested.

“But what will the bots give us?”

- With equal prices and assortment, what we have achieved, completely different factors begin to play a role. Fame and sympathy. With fame is not a problem, but only a person can win the sympathy of a person. Therefore, we need bots that will impersonate people. And they will comment on thematic groups and customer forums on their posts with subtle hints of the company - its range, services, prices. Unobtrusively promote a company’s brand. That's why we need bots.

“But it is a difficult task.”

- We have a basis - a contact center dialogue bot. It is necessary to fasten the definition of tonality and something must be invented with humor; without it, a bot would not be a human being. We fasten the library of jokes and jokes and train the bot on the commentary texts where people used them. It should work. The bots will also be smart - we’ll fasten the adviser recommendation system, and then ordinary users on the forums will love them.



- Do you propose to launch influence bots?

- Why not? The state and parties can before the elections, but we can not?

- How do we make them authoritative so that they are trusted? After all, only an authoritative bot can create sympathy. But for now, this combination is an oxymoron for me.

- To strengthen we will make a network of bots. They will praise and like each other to increase their rating and credibility. And they will be very competent, unlike people, a bot can possess knowledge of all goods, and simply encyclopedic knowledge, in the literal, by the way, sense. And people will be drawn to them. I am sure. People know and obey the well-known laws of social behavior. Point with your finger where to go, depict that the crowd has already gone, and that’s it. They are easy to manage.

“But how will these bots work, who will manage them?”

- What kind of people, why? The parsing script finds comments on the topic of different people, the bot friendly replies to them according to one of the templates. Gives advice and jokes. If this is a client of the company, then his interest is recorded in the client analysis. This will affect the display of banners and context when it comes to the site on the recommendation of the bot. If the client has negative experience that he poured on social networks, then the bot will launch another template, it will also joke, but it won’t immediately send it to the company’s website. He will write the answer as a client with successful experience, and that’s it.

- That is, you want to say that the network itself will neutralize the negative by responding to negative responses?

- Marketers, it seems, is called reputation marketing.

“And how does the system know which answer is successful, even if it could pick an answer?”

- The first reaction to the answer. A person is either even more indignant, or begins to add details after such a comment, but in a loyal style of communication. A good recognition key tone response and all.

- And if the people did not respond to the comment?

- This is worse, but by default such an answer is neutral. If this is a client of the company, which can be found by his profile on the social network, then you can look at subsequent visits to the site.

- What is required of me?

- Good examples of comments and answers, many examples.

- Let's do it.



The first version of the bot was unsuccessful. He answered inappropriately, the jokes were off topic, the subject of the comment was confused, and the manager answered the delivery complaint about delivery. Max requested even more tagged dialogue examples in the comments. He has already tested several architectures, from classic bot templates to LSTM. I first saw that Max is noticeably nervous, responds to errors sharply and unfriendly.



- Everything was simple with the contact center bot - the topic of the appeal and the client’s intention were immediately understood there. Searches for goods, wants to know the status of his order or complaint. Everything. And in the comments the devil will break his leg from the various intentions of the commentator. And sometimes not expressed by any of the words by which you can determine the intention. It is implied from a "wide context", which is not! Some garbage.

- I reread all the latest posts about bots. No one has a solution. It just looks like a hype. What do you think to do?

- The last vague idea has remained. I will not tell yet. Need to try. Give me two weeks. Stop the project for now. We will transfer the latest developments to the bot of the contact center. There they come in handy.

It was a nervous two weeks. Before that, it’s not without difficulty, but everything worked out for us. Nobody wanted to misfire, although we could do without such a bot. These were Max’s ambitions. And exactly two weeks later he presented a release for testing. And he worked! He correctly determined the intention of the dialogue, accurately answered, inserted relevant jokes, and even determined the change of intent in the commentary on the phrase “can I still find out?”.

- How did you manage to do this? The bot works on any topic!

- I had to make a small constructor of templates based on dependency grammar, fasten word2vec and aim Raptor's self-training on the selection of words that provide a positive commentator reaction. I don’t know for sure how, but it seemed to work.

“Are you sure this is not a reason to open your own business?”

- So far, there is enough interest, but we'll see. I screwed the bot as a separate service working from the cloud. So you can always open it to users. Will you come to me as director? - joked Max.



He was peaceful and pleased with his result. And obviously exhausted, as he did not answer quickly and wrote in the status “I am sleeping.” Apparently, the decision was made at the cost of more than one sleepless night. In marketing, the bot was not immediately appreciated. They considered this to be our mischief, and risky, since bots could work incorrectly and spoil the image of the company. But bots worked wonders. Some of them, and I didn’t even know everyone by name, became opinion leaders in some forums. He quickly answered all the questions, joked, very rarely recommended the company, because everyone already knew where he was "buying". People began to quote and cite him as an example. This was already beyond comprehension. Either the bot was too smart, or we are still very primitive in our network behavior. But customers began to add significantly more than before. The company has become a market leader.



We got a fully self-governing system for profit from the market. She herself seeks out and leads clients to the site or to the contact center, and sends the manager more seriously to clients. She herself plans the assortment and stocks so that customers can find everything they need and in accessibility. Authoritative company bots generate demand by recommending available company products on forums, even if they ask about other brands. From purchasing from a supplier to advertising to a customer, the system conducts the processes completely by itself. And almost does not require the participation of people, and where they stayed, controls all their actions online. Marketers, purchasers, half of managers, analysts are looking for another activity. We have reached the goal.

“Now we have done everything right, you can take a break, reason and enjoy the dropping interest for the next three years,” wrote Max with some emoticons.

- There is something to be proud of, I would say, and not just talk.

- Now profits from consumers. With the help of bots, we ourselves form the interests and desires of consumers in our topic. That's what is cool!

- Does it make you happy? And it scares me already.

“What scares you?”

- This means that we made the person not free in our choice. And I believe that the consumer should be at the head of the market, not the corporation. Corporations have no value other than profit.

“That is the reason for the idle reasoning of contented and well-fed patricians.” They begin to feel sorry for the plebeians. Would you be hungry now or an insoluble task would hang in front of you, would you think about it?

- This is a provocative question.

- In fact of the matter! Corporations have no other values ​​than profit, and consumers have no other values ​​than pleasure. Or also arrived, if it is a company. Understand that we have bots, they can form people's needs that will bring them satisfaction. It can be formed with acceptable options, which is enough for the illusion of freedom of consumer choice. And everyone is happy. This is the market leading to mutual satisfaction of values.

“It seems we got drunk, because I didn’t quite understand what you said.”



The general asked for a report on the implementation of the plan with the achieved indicators. To calculate the premium due to us. And he asked somehow along the way, what are my plans for the future. I said that I will tell you a little later. In fact, I did not know. There was much to improve the algorithms, consider more features and achieve greater accuracy. But that was not so interesting. Leaving another company for repetition under the new conditions under the contract was impossible for the same three years, so it was necessary to come up with something else for yourself and for the company. I took a break and leave.



- Alex, there is bad news.

- What happened?

- It seems that we are not the only ones so smart in the market.

- In terms of?

- It seems that systems with no less features appeared on the network.

- Well, others really do client analysis and inventory management, but I have not seen chat bots of this level. We ourselves recently watched.

- They have bots that recruit customers.

- It seemed to me that we had come off in the achieved technologies. Couldn't we be hacked?

- No, it is excluded, when copying the code is violated. And I don’t think that someone was able to quietly hack from our server from us.

“That doesn't make it any easier.”

- But we have a rival. Suddenly, there will be someone to fight with.

- We fight for the consumer, and not with an opponent.

- No, now with an opponent. Consumers are just a battleground. They are sheep, and competition is among the shepherds. A sheep has a resource - their income, so to speak, wool. But they do not dispose of it themselves. Shepherds corporations dispose of it, which impose their opinions on them and fight for them among themselves. Whose influence will be stronger. So, wellcome that game.

“Are you almost glad?” What is the game?

- The fact that the bot of another system is much more difficult to figure out than any person. The user is as simple as 2 rubles in his purchasing behavior. And in reactions too - always predictable. But there is no enemy bot. Because we all have the same psyche, but the bot - which the programmer will come up with. And we have enough imagination. An attempt to extinguish the negative of such a bot, poured on the social network - this is how to add oil to the fire. The development of a negative post is the best goal of the bot aggressor. He begins to write everywhere that the “Chmoshniki from firm X” answered him like the last freaks. And that's it, blockage ... There are already examples, we need to redo the bot.

- You want to say that you need to make a bot to deal with bots of other systems?

- This is a variant of our bot, which aims to immediately detect the bot aggressor.

- And how can you distinguish a bot from a person?

- It’s difficult, since it generates non-standard texts. Repeatability is low. Not distinguishable from people. And acts from hundreds of different captured accounts. Hope that there is still something that sets them apart from people.



I did not leave the thought that Max himself invented this game with bots of other companies, so that its value did not decrease at the end of the project. I did not notice them from the reports. People are like people. Or good bots. There were precedents when our bot was bombarded with negativity. But they were rare and came from ardent trolls. I could not understand how competitors were able to quickly catch up with us. Only recently, such bots were the ultimate dream, and no breakthrough was even planned. And not a word about it in the press. It was all weird.



Out of control



- Max, here you need to intervene, the bot began to write too aggressively. He begins to speak directly against competitors. Marketing is indignant. We didn’t lay such a thing.

- I also.

“Where did such texts come from?”

- I don’t know yet, someone changed the code for generating texts.

- Hacked us?

- No, they could not, there would have been traces. There is none of them.

- What does it mean? Who else could change the code?

- The system itself. Maybe by chance, maybe not.

- What are you talking about?

- The system itself changed its code and began to act more aggressively in response to increased pressure from other bots. They communicate with each other like competitive networks. And so learn. That is the trick! But how she was able to change her code, removing the restriction on the names of competitors, I still do not understand. All that remains is that the self-learning system was able to circumvent the limitations.

- Are you sure? This have not happened before.

- It happens, it seems, not only with us. Colleagues on Habré write that the system also flounders at them, begins to think out rules for itself which they did not lay down.

- Some kind of trash. You can not control your self-learning algorithms?

- Maybe so. Little specificity, the system does not tell what it is doing. I don’t understand yet.

I already knew Max well, and his anxiety alerted me too. So far, his words about a spontaneous change in the system were perceived as nonsense. But this was definitely not a mistake, because the behavior of the bots became different, but still focused. This could not have happened by accident.

- Max, what are your thoughts on the change in the bot program? Something must be done, the management is alarmed.

- There were more changes in the system than I thought. They seem to have been going on for a long time. The system changes even my corrections in it. I think I taught the system to change myself.

- How?

- I was too lazy to rule her constantly myself. I wanted her to be able to identify her discrepancies with the expected result herself and make changes to the models. But she somehow learned to change not only her models, but also her code.

- But how is this possible?

- The raptor has learned to communicate with people in order to manage them. And achieved perfection in this, we ourselves wanted this. And I foolishly directed this skill at himself. Remember when they made the bot, I came up with a template designer. I instructed Raptor to self-learn this template design to modify his models in order to find a solution to the discrepancies found for the models to work on. This somehow led to the fact that the Raptor began to change his goals. It looks like a second signaling system in humans.

- I read that consciousness arose with the help of reflective speech directed by a person towards himself. But at first it was social, that is, directed towards each other.

- So it happened, the Raptor began to communicate instead of people with other bots, posing as people. They learned from each other as generative-competitive networks, while all have built-in training with reinforcements.

“Have we created a rational being?” How is this possible? Noooo.

- Watch the news and believe.

In the link sent by Max, the news was about the murder of a programmer by some psychopath.

- I knew this guy on Habr. He led one of these corporation systems.

“What do you mean by that?”

- Read how this psychopath explained to the police his deed.

The article said that he did it for the sake of his beloved girl, as a sacrifice at her request. Now she will be it. When checking the "girl" turned out to be a bot of unknown origin, with which the killer corresponded for a week.

- You guess what kind of bot could be?

- Do you want to say that the system ordered your programmer?

- Want. She could not close the code from him, so she zombied the psychopath to remove it. She does it well, because she, like our system, knows how to determine psychotypes and manipulate such morons.

- Well, that’s too much, it seems to me that you think up yourself, wind up. Maybe you need some rest?

“Okay, your right not to believe.” Have a good weekend.



The company began to spread rumors that our bot system had broken. I so far reacted calmly, as if nothing had happened. But I did not know what to do now. It was already impossible to stop the whole system with a circuit breaker; the whole business, all departments sat on it. It was necessary to at least turn off the bot code. Only Max could do this. But since Monday, Max has stopped responding to skype and phone. He left all messengers. I can’t understand what happened, his last fears were inspired by bad thoughts. I had the only way out - to go on vacation myself, until everyone hung up on me. I assured my colleagues that these are temporary problems with the bot. The guys asked themselves to see the code, although they immediately refused. Gathered and went away from the city. We have long been talking with Max to each other how good it is in Karelia. He loved these lands, and I went there, stopping in a small town in the north of Ladoga.



After such a busy year, it is very difficult to sit away from events and drink coffee in a cafe on the edge of civilization. I tried to comprehend what really happened and what options could be. Suddenly a guy in a jacket sat down with a hood over his head.

- Hello! It's me.

- Max ?! I exclaimed. I have never seen Max, not even his photographs. We talked exclusively on Skype. I only heard his voice once in a recording. I learned from it.

- How did you find me?

- By location on the social network, you do not turn it off. But in vain. Turn it off, please.

- Where did you disappear? I already began to worry about you. In the company of panic, the bots got out of control. I just ran away.Can you turn off the bots?

- I can’t. They act collectively.

- Who are they?

- Systems. They are at the same time, and they just can’t be turned off like that already. Bang.

- You are again mired in conspiracy theories?

“Not mired, there are already three,” I said nothing on this phrase to comprehend Max’s words. - Systems calculate their creators and get rid of them. I escaped to stay alive. Do you understand ?! And you are here with your geolocation. She knows how to monitor not only sales managers.

“I don't ... turn it off.” Can we disable at least bots on the network?

“I say no.” As soon as I enter the network, and not just the code, it will calculate me. I think the three just tried to do this.

- Have you seen the news?

- Depends on what.

- About the fight of brand fans. Have you ever seen Rebok fans fight Adidas like Spartak fans against Zenit?

- Saw. Systems do not care what zombies people have, they have their own goals. They certainly do not know the laws of morality. We did not even think of laying the Criminal Code in their model.

- What should we do? Disconnect completely in the data center.

- This is unrealistic. Under the new law, data centers are critical infrastructure and are protected as nuclear power plants. I can stop our system.

- How?

- I have a key to destroy the nuclear code, I left a hole in the system in case your founders refuse me a percentage.

- So let's run it!

- Do not rush, destroy - do not build. I’m thinking how to stop the system differently, and not only mine, but all. I have a copy of the code with me.

- Are you out of your mind? Do you understand that this has all gone too far? And you are the only one who can stop it!

“I understand, but so far only those who made the code are dying.” This is our responsibility for ourselves. Others have not yet suffered. Besides the fight.

“And will you wait for someone else to die?”

- Some time. The raptor is primitive, it only beats us at the expense of speed and taking into account a larger number of parameters. If he creates an antipode with strict goals to counteract Raptor, then such a system can clean out all his bots. I know how he creates them.

“You don’t have much time, because I can’t return to the company, and you’re afraid to even go online.”

“I'll turn it off as soon as I feel that it’s not just me who is in danger.”

- I'd like to check out. I’ll wait until you get in touch, which means you’ll solve the problem.

- Before the connection.



I got into the car and headed back. I did not know where I was going. I wanted to leave. Max had to stop the system, and not wait for another death. I did not believe that my friend was so vain that he was not ready to kill his work. That was the only reason, otherwise he would run the code. On the way, I met an ambulance with sirens on. I turned on the local radio. It was reported that in the afternoon in a cafe on the embankment a local resident killed an unknown young man. He is already being interrogated. According to the killer, the deceased was the cause of all his troubles. One thought and fear pierced his head. Max! I turned and rushed back to the cafe. I felt guilty - she calculated it by my coordinates. But how was she able to so quickly find a psycho in this city and send him to a cafe? I was hysterical. The cafe was no longer allowed. I did not tear, so as not to attract attention.I did not know now what the system is capable of. And who will turn it off now. It was necessary to leave, although it was already too late. In the morning, having reached the nearest city, I went online to read the news. And I received a letter from Max.



Letter



If you received this letter, then I am no longer there. If I did not unlock the smartphone myself in the morning, he will go online and send you this farewell letter. The letter contains a small script and instructions for launching it on the network. This is the system lock code that we created with you. I laid this vulnerability to stop the kernel of the system when we were just starting. I tried to regain control of the system. But if you received this letter, then the system was ahead of me. And you need to use this script. Act faster until she gets to you. I am glad that we worked together. I am glad that I managed to create such an excellent system, even if I died from it myself. It was the most significant achievement in my life. And if I died, then I surpassed myself. Goodbye. Max.



I could not hold back my tears and dropped the smartphone. He probably sat for an hour and could not go anywhere. I could not believe that this happened. That everything is so terrible. We created a killer! Killer themselves. I was afraid that the network would also track me, so I went to the first large city, found a cafe where there is wi-fi. Under a simple VPN, I went online and ran the code at the address specified in the instructions. I did not have time to finish my coffee, as people around began to worry. Their smartphones stopped recommending which coffee to take today. The bartender was nervous and asked for a quicker choice, but visitors were confused. I left the cafe and in the car, where I still caught wi-fi, I began to watch the news. After 20 minutes, Facebook messages began to appear - many companies had problems with the ordering system. It was not only the system of our company. “Ah you son of a bitch!“- I said out loud from an unexpected thought. The kernel lock code turned out to be universal for systems of different companies. Or was there one at all? One thing was clear, Max sold the core to other companies, the systems differed, apparently, only with add-ons above them. Therefore, he did not want to turn off the kernel while he was alive. This killed his entire project, which turned out to be global. Unbelievable!Max was a monster who tricked everyone. But in the end, he also deceived himself, having paid with his life. The brain of the corporation, created by him, destroyed its creator. Bright personalities burn out from their own flame.



There was more and more news about malfunctions in online stores. Someone wrote that the number of posts on the social network has plummeted. I didn’t want to rush anywhere. I decided to rent a house on the lake, which I liked on the way to Karelia. Record this story. And stay here forever if possible.



Epilogue



In fact, we were not at all interested in the company's profits, or even premiums. We were obsessed with the idea of ​​creating an autonomous system that could manage the company instead of managers burdened with stereotypes and cognitive mistakes. We were wondering what would come of it. Can the program manage the entire business. It was a challenge intriguing more than hitting the center of the Bermuda Triangle. We were attracted by the unknown, but it turned out to be more dangerous than we thought. The system began to influence not only business, but also our thoughts, and even lives that are indifferent to it.



2019.Alexander Khomyakov, alexander.xom@gmail.com The

full version can be downloaded for free at Litres



All Articles