When AI is used for evil

With advances in artificial intelligence (AI), the risks of hackers using those technologies to unleash cruel attacks are increasing, leading researchers warn in a public report. Dad on February 21, 2018.

They can turn consumer drones and self-propelled vehicles into potential weapons, for example, according to researchers from universities like Oxford, Cambridge and Yale, as well as teams organized by Elon Musk.

The report 'Using artificial intelligence for evil' warns of various security threats when taking advantage of AI. For example, self-propelled vehicles can be tricked into misinterpreting a stop signal, which can cause traffic accidents, while a bunch of unmanned aircraft are controlled by an AI system, which can be used to monitoring or launching rapid and intermittent attacks, the report said. Smart machines, according to reports, can lower the costs of conducting cyber attacks by automating certain tasks that are labor intensive and more effective in identifying goals. potential.

AI may be the worst event in our civilized history.

An example of a report indicates a "phishing attack" , whereby attackers use personal messages on each potential target, to steal sensitive information or money. 'If some studies are appropriate and synthetic jobs can be automated, many people may engage in phishing attacks , ' the researchers said.

On the political front, AI can be used to monitor, create more targeted propaganda and spread false information. For example, "high-definition videos " show images of national leaders making provocative comments they never did. 'We also forecast new attacks thanks to the progress of our ability to analyze human behavior, mood, and beliefs based on available data' , reports data. AI has been exploited to create fake images, overlapping others with video.

For example, a video called 'deepfake' cradles a human face in adult porn movies. Recently large websites have begun to practice this. Experts believe that the scenarios highlighted in the report are not accurate predictions of how harmful AI is being used - some scenarios may occur over the next five years, while those that may occur in the next five years, while Other scenarios took place at a more or less limited level. 'Other malicious uses will undoubtedly be discovered that we did not predict , ' the report said.

The February 21 report did not provide any specific way to prevent the use of evil AI. But only provide some recommendations such as calling for more cooperation between policymakers and researchers, and calling for the involvement of many stakeholders to address the misuse of AI. Although technology is still young, billions of dollars have been spent on developing AI systems.

International Data Corporation last year predicted that by 2021, global spending on cognitive systems and artificial intelligence could reach 57.6 billion USD. AI is predicted to be so great that Google CEO Sundar Pichai recently said that it will have a deeper impact than electricity or fire - two of the most popular innovations in history.

At the same time, there are many skeptics about AI. Renowned physicist Stephen Hawking said last year that the AI ​​could be 'the worst event in our civilized history' , unless society finds a way to control that development.