Influential Researchers Sound Warning on Threats Posed by AI

PHOTO SOURCE: MOTHERBOARD.VICE.COM

A group of 26 top AI researchers from institutions around the world just issued a 100-page report detailing the digital, cyber, political and physical threats resulting from humanity's increasing dependence on AI.

The report was compiled after a 2-day workshop held last year in Oxford, UK. Below is a summary of the threats highlighted in the report.

Digital Threats:

• Misuse of machine learning tools (The recent use of a machine learning library - TensorFlow - to create fake porn by superimposing the faces of Hollywood stars onto the bodies of porn actresses.)

Cyber Threats:

• Large scale phishing and hacking of security systems (Criminals may harness AI to dramatically scale up phishing attacks and increasing the efficiency of cyber attacks on the security systems of governments and corporations.)

Political Threats:

• This is linked to the digital threat posed by machine learning tools. Villainous parties seeking to inflame existing geopolitical tensions to serve their own agenda may air a fake video of an inflammatory speech by a politician.

Physical Threats:

• Vulnerability of self-driving cars to sabotage (Malicious and tech savvy individuals could easily utilise AI to impair the systems in self-driving cars and cause deadly accidents.)

In light of the above clear and present threats posed by the advent of the AI age, the researchers strongly urge the development of policies to "ensure that AI technologies don’t fall into the wrong hands, or instituting some sort of monitoring program to keep tabs on the use of AI resources, " according to Motherboard.vice.com