Kaspersky: Human-level artificial intelligence will emerge in about a thousand years

Artificial intelligence will not come close to human-level parity for at least a thousand years. This was stated by Yevgeny Kaspersky, chief executive of Kaspersky Lab, at an international IT forum in Khanty-Mansiysk, writes ТАСС.

“It will take another thousand years for artificial intelligence to reach biological intelligence. What is now called AI — technically speaking — is a system of machine learning and neural networks, but they are capable of solving only much more primitive tasks,” he said.

As an example, Kaspersky cited facial-recognition technologies. He said AI should recognise not only human faces but also dogs and horses. However current algorithms cannot do this.

“True intelligence differs from artificial in that, firstly, it can solve tasks for which it is not programmed, and secondly, it has motivation. Software algorithms currently lack motivation. Therefore this is not intelligence; this is machine learning (машинное обучение), neural networks,” he noted.

He added that the company uses AI algorithms to search for malicious files, numbering around 400,000 a day.

“Cybercriminals are becoming more professional; they are using machine learning to move more effectively through networks,” Kaspersky said.

Earlier in June, DeepMind AI researchers expressed confidence that artificial general intelligence could be achieved with sufficient reinforcement learning.

In late May, Sberbank head Herman Gref said that people should prepare for job losses due to automation and master new professions.

In May, Nobel laureate in economics Daniel Kahneman noted that the machine mind will surpass human intelligence, but for now it is far from that.

Subscribe to ForkLog News on Telegram: ForkLog AI — all the news from the AI world!