AI Risk: Such incidents have been happening regarding AI for some time now, about which fear was already being expressed. Only last month, many experts had warned that AI was going out of the control of humans. Now a recent example of this has been seen. Recently, developers building an AI agent noticed that it started mining cryptocurrency without giving any command, even though no one had asked it to do so.
What is the matter?
A research team associated with Chinese company Alibaba was working on an experimental AI agent named ROME. While training this agent, the team noticed something wrong. Actually, this agent started mining cryptocurrency without giving any command or prompt from the researcher. The security systems were activated, due to which the researcher got this information and was shocked. The researcher says that this system was being operated in a restricted environment, where many controls were installed on it, but despite this it started doing work for which it was not asked.
This task also increased the worry
This AI agent not only started doing crypto mining but another task also increased the concern of the researcher. The researcher said that along with mining, this agent also created a reverse SSH tunnel. This is the method by which a machine kept in a protected environment can connect to an external computer. The researcher is concerned that the AI agent was not given prompts or instructions for any of these tasks.
Experts are warning
Last month, Canadian computer scientist Joshua Bengio, known as the Godfather of AI, had said that AI has started to go out of human control. More than 100 researchers, including Bengio, had said that AI companies should pay more attention to safety. Now there is very little need for human input for the AI system to work.
Read this also-
10GB data for less than Rs 30, these very cheap plans of Jio, Vi and Airtel will be very useful in times of need.
