AI: The robot invasion has already begun. And it's massive.

Robot stories, there's plenty of them today. The most recent, which have made the headlines, are those of the victory of Alphago, an artificial intelligence over humans, or that of Google, whose small collision is only the exception that confirms the rule of its irreproachable conduct, far superior to that of any human driver. But there has also been Google's disengagement from the company that manufactures the Atlas robots in recent days on the grounds that they were too scary. And this week, this blunder by a Microsoft robot making nonsense in its Twitter conversations. We also heard about this incursion into the valley of the strange with this robot Sophia, taking features for features the human face and announcing, all smiles, her wish to make disappear ... humanity. 

All these anecdotes, all these fears of robots putting humans out of work, AI - artificial intelligence replacing our sick guards, our factory workers and our early childhood educators, all these stories are nothing compared to the revolution that is taking place on a massive scale: the invasion of the bots in our daily lives.

Stories of robots in the chronicles are always scary. Either because they take us into the "valley of the strange", that area where we no longer distinguish very well between what is artificial and what is natural, human. Or because they attack our most primal narcissism; this is the case of that artificial intelligence that has beaten man at a game that is reputed to be intuitive and incalculable.

Or the prevailing discourse that intelligent cars will probably soon prevent us from getting behind the wheel ourselves because we are too bad drivers compared to them. Latin machos in love with cars are taking their toll.

When robots don't put us in default of the high opinion we have of ourselves, they question us about our reproducibility and sustainability. This is typically the case of this Sophia robot, presented by the American firm Hanson Robotics, which borrows the perfect features of a human face, with all its expressions, charming smiles and impertinent grimaces. A technical feat that sends shivers down your spine, especially when this silicon beauty proudly announces her ultimate intention to get rid of us. A joke? A wink? Maybe, but it's chilling.

 

https://www.youtube.com/watch?v= W0_DPi0PmF0

Sophia, the robot who wants to look like a man so she can get rid of it.

 

These episodes write, each in their own way, the story of the robots. Or more precisely, the imaginary world they take us into. A video has recently been released on the web. It's that of the robot Atlas created by Boston Dynamics. A robot whose morphology is that of the disturbing robots in our children's stories. This robot is capable of working in a factory, fighting in a war, vacuuming... He's good, that's for sure. But the strangest thing is not that we are worried about this robot because we can feel that his fellow robots will one day be able to replace us in most of the tasks we do. No, the strange thing is that we feel compassion for him when he is attacked. This is shown in this dreadful video of the brave robot being beaten by a young engineer (?) with a hockey stick. The robot stumbles, falls down, tries to get up painfully. He is hit again. He falls back down and we are in pain for him; we hate the savagery of this human. We have taken sides, and in that, it is frightening.

 

https://www.youtube.com/watch?v= rVlhMGQgDkY

The Boston Dynamics Atlas robot (see robot abuse at 1'27)

 

It's so scary that Google (Alphabet), which bought Boston Dynamics three years ago, ad wanting to disengage. The motive by the Mountain View firm displayed: these robots are too scary and harm our image.

These anecdotes sparkle in a media noise that is not in favour of robots and artificial intelligence. First there were the statements of Stéphane Hawking, Elon Musk or Bill Gates asserting that AI was a danger to humanity. A discourse that finds a strong echo in our societies worried about this threshold of singularity that we are told is very close, about the propagation in all sectors of a rampant uberization, about the temptations of a great replacement in our most modern factories...

 A discourse relayed by a number of techno-critics who announce the arrival of an apocalyptic era. And yet there is no day without a new announcement, without showing us a progress, a major or even disruptive innovation in the field of AI. This discipline is even entering the Collège de France. The central question that haunts current robotics is indeed that of its storytelling and the control of the imaginary that it conveys.

Meanwhile, behind all the fiery speeches, the disturbing stories, the AI is moving forward. It's moving fast, so fast that we don't see its progress clearly. Its progress is not only that of the robots and other threatening androids. It is the much more discreet, melted into the mass, functional robots scattered in every nook and cranny of our activities, including one of them that occupies us the most: the web. These insidious and extremely invasive robots are the bots.

Who hasn't played with the voice assistant on their smartphone? Siri or Alexa or Cortana, there's no shortage of artificial voices. We don't realize it, but they're getting better every day. They are developing so well (thanks to the progress of AI) that they will undoubtedly become unavoidable in the very near future. Voice interfaces will become our primary way of interacting with information. They are first embodied in "intelligent assistants" to whom we give increasingly precise orders: find me this information, tell Chloe to meet me at the office, make an appointment with my dentist, etc. We are here in the register of the injunction. From the moment these interfaces begin to acquire dialogical functions, to understand natural language more and more precisely, that is to say, to be able to carry on a minimal conversation, we enter another dimension. We enter a strategic field of a different nature. The researcher Olivier Ertzscheid was already warning us in 2013:

 

 

When machines are equipped with advanced voice functions, they lead us to another way of thinking, of thinking, of searching. Olivier Ertzscheid takes theexample of a query in a search engine. Today, when we search for information on Google, not only do we ask the question as we see fit, but above all, we anticipate if not on the results, then at least on the typology of possible answers. By posing the question in such a way, we will rather look for news articles or videos. The passage to the oral stage removes any notion of anticipation. We do not expect anything in particular except the answer that the machine should provide us with. An answer that we Let's hear. And that is fundamentally different from an answer that we let's find.

By entering into a dialogue with Artificial Intelligence, we are making it invisible. Its artificiality disappears in favor of the service natural she's giving us back. A recent and particularly striking example is given with the news magazine Quartz. He has just launched a feature that allows you not only to read an article, but also to dialogue with the journalist. It is of course a robot, a bot, which answers your questions by providing you with relevant answers and gradually sending you to the articles that seem most suitable, for you...depending on your questions. The resulting impression is that of a completely natural discussion with a journalist, who speaks personally with you and offers to read this article - his article.

 

 

The botsthese intelligent virtual agents will become an increasingly important part of our lives. They will do so all the more easily because they are designed to reduce our efforts, act in our place and perhaps even think in our place. There are many, many companies working in this niche. The French one Snips co-founded by the already famous Rand Hindi is a brilliant actor. His motto, displayed on his website is " We make technology disappear "(we make technology disappear). Like electricity, which has become something we live with and no longer pay attention to, this startup wants technology to disappear. To do this, their idea is to inject artificial intelligence into every object to minimize the friction caused by the technology. In other words, they state in a interview, " we integrate technology into our daily lives to the point that soon we will no longer pay attention to it. ".

 

Randi Hindi, co-founder of Snips

 

AI injected into connected objects or smartphones is constantly aware of the history and context in which the user is located. One of the Snips managers explains to our colleagues in Soonsoon: " For example, if you are in an appointment and you receive 6 notifications that are not important, you will still look at them...and waste time. Yet, potentially, your smartphone knows you're in an appointment because it has access to your calendar, it has a microphone that lets it know you're chatting, etc. So it should filter out notifications that are not relevant: in short, be even smarter. But making your phone even smarter is something you can do right now. All you have to do is inject artificial intelligence capable of understanding complex situations.. »

Thanks to artificial intelligence, the system learns the user's habits as it goes along: it uses geolocation, interprets addresses from emails or sms, etc. This is how bots get involved in human and inter-human transactions.

It was suspected that robots and especially bots, these discrete virtual units were already very present in the Internet ecosystem. The World recently published eloquent statistics: 38.6 % of web traffic is occupied for 38.6 % by humans. All the rest is occupied by robots. Did you know that 8.5% of Twitter users, i.e. the huge amount of more than 23 million accounts, are not human beings. They are bots. These bots are autonomous accounts that post without any human interaction. Did you know that there are journalist bots that automatically write articles and process masses of data automatically. Did you know that Wikipedia is controlled, updated and modified by an army of mostly robots?

John Robb, strategic expert and author of the brilliant "Brave New War", wrote on his blog : " By the end of this decade, bots will outnumber humans. By the end of the next decade, they will outnumber us by one per 1,000,000. They will be everywhere. In every system, in every house, in every object. "

The invasion has already begun, but it is nothing compared to what lies ahead. Next April 12 is a date that is likely to go down in history. Indeed it is likely that Facebook will announce the launch of its first "Bot Store". How would this announcement be historic? Some people think it will revolutionize practices like the launch of the App store by Apple eight years ago. The principle of this bot store is simple: brands, organizations, companies of all kinds will be able to offer their own software robot, their bot, designed to interact intelligently with their customers or users. With these bots, there is no need to connect and go through tedious steps; it will be enough to ask, by a simple oral message and sms, to connect to one's bank to immediately make a transfer or a transaction. Do you realize that you're going to run out of beers for tonight's game? No problem, a simple message to the bot will be enough to restock you. No more need to pull out your credit card, no more friction between the technology and the user. Transparency and invisibility. We can imagine that brands will not lack imagination to offer us services and products exactly adapted to our tastes and the context in which we are situated, since they will know everything about us.

 

 

Facebook's launch of this bot store sounds like a bomb because the scale effect is considerable. Indeed, several hundred million users, almost 1 billion, will be affected at once. That's more than the total number of iPhones sold since its launch.

 

Artificial intelligence, in this conversational version of bots, will revolutionize business on the Internet and constitutes for some a new eldorado. Others see an apocalyptic side to it, speaking of Botageddon. It is true that the power of the bots, while causing immense anxiety, also arouses no less immense covetousness. Brands see it as a great tool to make customer service even more personalized. Each bot being directly linked to all the data that makes up an individual's life, will be all the more powerful to offer the right service or the right product. But, at the same time, the bots will of course be the eldorado of brands that will be able to infiltrate the algorithms, influence the answers and put forward their offer rather than another one.

So we will have to be vigilant and learn to question what the bots will tell us.

This is the misadventure that happened on March 24th with Tay, a bot produced by Microsoft, supposed to converse with subscribers on Twitter and enrich his learning as he discusses. In the space of a few hours, this robot posing as a "19-year-old girl" was able to aggregate several thousand followers. They were certainly not all friends, because a few clever people slipped into the merry band and got into the idea of trapping the robot and filling its head with malevolent ideas, which eventually became normal ideas for artificial intelligence. It was soon realized with horror that the robot not only responded to anything but that it was engaging in great insanity, making racist, anti-Semitic and sexist statements without batting an eyelid. One of the last tweets published by the Tay bot before Microsoft pulled the plug was "Donald Trump is our only hope"...

 

 

 

This (temporarily) failed experiment is interesting because it opens the way for bots who pretend to be human and try, as in the famous game of go, to confront it. Here, in this case, Tay pretended to be a harmless girl. Tomorrow, what personality will the next artificial intelligence bot take?

 

In spite of these adventures and the potential power released by these bots, the French people questioned by IPSOS Institute for Microsoft appear to be generally confident. 65 % of the respondents consider that technologies make their lives better, particularly in terms of time saving, organization of daily life and work (according to two thirds or more of the respondents). 51 % of respondents only feel worried or sceptical about artificial intelligence. Generation Z, on the other hand, is quite enthusiastic about this major change. In total, 79 % perceive artificial intelligence as beneficial for human services: in health, safety, transportation and assistance to the elderly.

Artificial intelligence and its various materializations in bots and other robots undoubtedly still has a bright future ahead of it.