The threat of “machine uprising” has been described  by futurologists since the very concept of artificial intelligence emerged . You can find such references in the books of King and Bradbury, and in many films like “The Matrix”. But is AI indeed as dangerous as they say? Stephen Hawking shared his concerns about artificial intelligence: “The development of autonomous artificial intelligence can put an end to the existence of the human race.”  He also mentioned that automation brought by AI would accelerate the already growing economic disparities around the world and would allow small groups of people to make huge profits using very few assistants. And such fears do not arise from scratch. For example, the American company Goldman Sachs has already replaced traders involved in stock trading on behalf of large bank customers with an automatically running AI-based bot. Now, out of 600 people who worked in the year 2000, two are left – the rest have been replaced by trading robots.

      Elon Musk  thinks AI is one of the biggest threats to humanity

Despite these concerns, developments in the field of artificial intelligence  are in full swing. Large researchers in the field of development and creation of computer technology invest money, effort and time in teaching computers to behave like people. Some have already succeeded in this field. Machines can understand people, talk with them, or even learn from them.

Amelia

It may take time to create a perfect machine that will be able to  communicate with a person freely, but a basic version for business has already been created. New York-based software development firm IPsoft has created an “employee with cognitive knowledge” called Amelia. Together with Accenture, the company works to develop a pilot software for two large oil companies, Baker Hughes and Shell. Amelia has anthropomorphic features.

 In terms of perception it corresponds to a 6-year-old child. Amelia has an emotional spectrum through which it can respond differently to the tone of the interlocutor, explained IPsoft employee Martin Gribnou. If the interlocutor complains, Amelia respectively changes the tone of her voice. If she cannot solve a problem, she calls a human operator, observes his actions, and learns. When a person types a message to her, the software breaks it up into parts. During the conversation, Amelia can correlate current questions with previous ones, complementing the answers.

Initially, Baker Hughes was going to use Amelia to work with external suppliers; she had to handle requests, invoices, etc. At Shell, Amelia will help to develop new advanced training courses for employees, said Cyril Bataller of Accenture.

 

Yseop

French company Yseop (pronounced “iz-op”) develops software of a different kind. Yseop can create personalized reports and recommendations based on the dynamic data of daily users, says company director John Rauscher. On websites such as vetonline.com and L’Oreal, this software asks questions of users and generates answers in simple English, just like a doctor or a stylist. In Yseop, the principle of AI called the “logical inference mechanism” is involved. Using the example of customer relationship management software, Rauscher demonstrated other features of the developed program, such as creating short and easy-to-read reports. The bottom line is that nowadays information is easy to find and store, but processing data deposits is a difficult task for humans. A machine can help a person by quickly processing a large amount of data and summarizing it in plain language. This is only the beginning of Amelia and Yseop. These are rather primitive forms of AI, and hardly any of them will have a plan to take over the world.

Among their main important functions is working with repetitive tasks. Call center employees follow approximately the same routine, be it an invoice or a pet health question, which can be stressful for an individual since creative tasks are reduced to zero. Machines do a good job with such types of assignments.

With other professions that require intensive work with information and knowledge, AI will help people, rather than replace them. For example, sales professionals no longer have to process tons of information to generate a small report on their customers. Doctors, lawyers, employees of the financial sector — everyone who deals with large volumes of data — can send inquiries to such assistants as Amelia and quickly get an answer about regulatory acts or medical history. Ambulance workers can automate the admission of low-risk patients using Yseop software, which will ask standard questions and create short reports based on the answers. This is an extremely useful and necessary use case of artificial intelligence applications that can save people lots of time and energy. At this point they act as an extension of human capabilities rather than a threat to their well being.

Clearly, concerns about “machines threats “ are premature. AI powered systems are capable of solving only certain tasks. They cannot surpass the human mind.

The emergence of super intelligent systems should be expected no earlier than 2045, the American futurologist Raymond Kurzweil predicts. But according to Microsoft co-founder Paul Allen, until the human brain is thoroughly studied, it’s too early to talk about a potentially dangerous supermind. We are very far from the level of advancement of the machines that can actually become smart enough to set up their own goals and compete against humans.