AI has many implications for society. Most are good, but some are bad. For instance, some are predicting that AI can soon be used to commit smarter crimes. Though using software to commit crimes is nothing new, it could become a larger problem when AI opens up new possibilities for the commission of crimes.
For example, imagine if a computer could create a synthesized voice that sounds just like a close relative? Imagine the potential for crime when a computer can pose as your mother or grandmother on the phone to request sensitive information such as banking login information or bank account numbers? Though it may sound like science fiction, it isn’t exactly far-fetched. AI companies are already working on AI programs that can mimic human speech that sounds surprisingly natural. In just a few years such technology may be readily available and affordable for anyone who wants to take advantage of it to commit crime.
This is just one example of how AI could be used to commit a crime. AI can also be used to create more sinister malware that could wreak havoc on our current antivirus programs.
In 2003, researchers at Carnegie Mellon University created Captcha—a kind of Turing Test that prevents computers from stealing online accounts. Now programs are being developed that are smart enough to fool Captcha and essentially pass for humans in order to gain access to sensitive account information.
If there’s any consolation, it’s that law enforcement and cyber-security companies will have the same technologies available to combat such crimes. Better cyber-security and improved monitoring systems will help law enforcement and cyber-security companies to keep up with smarter cyber-criminals.
The potential for increased crime exists as AI is developed. Artificial intelligence companies and especially machine learning startups, need to anticipate this and prepare accordingly, because there’s going to be a big market for such technologies.
Source: The New York Times