Assistant instead of Skynet: what kind of artificial intelligence does the state need

The state and business seem to have found real use for machine learning and artificial intelligence (AI) technologies. AI is able to solve complex, but rather narrow tasks, removing routine tasks from a person. The main thesis is that the machine does not command, it becomes a digital assistant and needs a teacher.

Back in 2017, Gartner noted that the publicity around AI overshadows the real benefits of technology. In the wake of fast investment money, companies and start-ups were carried away by semi-fantastic developments. AI systems are being delegated tasks that humans are much better at today.

I am sure that business and the state should pay more attention to AI technologies that are already working well. Among them are decision support systems (digital assistants) and business process automation.

For example, AI automation systems for oil refineries (refineries) have demonstrated their effectiveness. Based on various parameters - the conjuncture of supply and demand, etc. - the machine itself predicts what is more profitable to produce from the incoming grade of oil . She rebuilds the production process on the go, thus solving a very important, but purely professional issue.

When introducing AI into public services, difficulties of a special kind arise. AI projects are usually associated with testing hypotheses: the result can be negative or philosophical. For a long time, plans for the implementation of AI projects under 44 and 223-FZ were broken against this contradiction. Government customers need proven effectiveness of the implemented solutions, otherwise there is a risk of inciting a disaster called "misappropriation of budget funds."