For most people, though there is a growing concern over malware, viruses, and cyberattacks, the more complex issue of artificial intelligence (AI) cyberattacks is something we leave to the “experts.” Most people only understand the basics of AI, but those truly in the know see the potential for a very costly wave of attacks on companies and individual users. It is something that we are aware is in the realm of possibility, yet something that is outside of our control.
Whether we want to admit it or not, we have created this genie who now roams across international borders and is able to play havoc with key infrastructure systems and the basic fabric of our daily lives. We complain when our cell network goes down or the power goes out in our home. But all of this is connected and controlled by computer systems, systems that are under constant attack by enemies foreign and domestic. Not being able to access the Internet is one thing; not having any electricity at all will cost us more than just time and money.
The AI technologies being used to attack our basic systems learn as they go at a far faster rate than we can comprehend. Every piece of data adds to the strength of the AI attacker, as it seeks out new ways and new places to attack. As long as there is power to keep the AI programs running, they will continue to learn and create new methods of attack. That requires an AI defensive presence up to the task of countering the attackers, and must be done as constantly as the attackers adjust their tactics.
If you feel this is out of your control, you are not alone. The problem is that first world countries have largely given control of their financial, military, and infrastructure systems to computers, and so are at their mercy. Mercy may not be a good word here as AI systems simply carry out orders without first seeking the approval of their creators. This is how they were designed, this is how they were programmed. If any AI system cannot see the end of its actions, such as causing an airplane to lose complete power at 50,000 feet and plunge 200 people to their deaths, it has no capacity to consider the morality of the decision.
What is often overlooked in this arena is the inextricable connection between computer and electricity. We need electricity to perform most of our basic daily tasks (refrigeration, heating, cable, Internet, lighting) and computers need electricity to run, but because we are so tied into computer systems we can’t have electricity without allowing computers to operate. So AI learning will go on as long as we maintain our dependence on computer systems. This means that AI cyberattacks are just as much a part of our daily lives as turning on the TV or radio to get the morning news.
It was mentioned earlier that the cost of these AI cyberattacks go beyond time and money. For most people alive today, it is impossible to envision a fully functioning, normal, productive society without electricity or computers. The unsettling part of this reality is we created the genie but had no foresight into how to get it back into the bottle. The most vulnerable countries are those whose culture is most dependent on computer systems. Those countries will have to turn the defense of their way of life over to AI creations that are able to defend against the ever growing, more complex AI cyberattacks.
But it is not human programmer versus human programmer. It has been suggested that a time will come when human programmers are no longer needed because computers will create their own programs. It is unlikely that defense was going to be the first area affected by this reality. There is an irony as many of the new jobs that are being created are jobs that require creativity. Our own creations are telling us what to do. Choosing otherwise can result in a catastrophe.
We perform normal actions every day that we don’t give a second thought to, but the AI programs are collecting, storing, analyzing, and calculating their next move based on our past actions. Facial recognition, biometric, and password information is all stored on what are supposed to be secure servers. But secure from who? The perception is that they data is secure from people but accessible to computers for storage, retrieval, and analysis at a later date. Where do the humans enter the loop? The 21st century answer is, they don’t. Computer systems know all, and AI is an assistant to extract information on everything from our eating habits to our personal preferences on movies and music.
A logical next step is for AI systems to make virtual copies of people. The only way you know what distinguishes one person from another in the digital world is their digital information. Computers are able to understand and create voice patterns, so why not imitate your voice or anyone else’s? Writing style or signatures are identifiable, and combined with the database of a person’s likes and dislikes, how is a person to distinguish between what an AI system has created and that of another person?
The classic movie Wargames is being remade. The original was made in 1983 when dial up modems and 64k connections were considered fast ways to exchange data with a computer system. The world was saved because a geeky teenager told the WOPR to learn from its mistakes. The world of AI is not constructed to take orders from its creators, but to make its own judgments and take corrective action.