AI technology is not perfect
With all the practical benefits of AI aside, there is a serious issue that needs to be addressed. It is the issue that many top scientists, business leaders, and others have been talking about for a few years.
AI can be used for nefarious schemes and purposes. Its double-edged sword allows many people access to top secret facilities by simply designing harmless ads on social media sites. Well, that is one of the scenarios that come up in these discussions.
Other scenarios depict harmless looking AI robots to become robot assassin. They use different technologies to recognize their targets and detonate bombs. In other words, a robot kills President.
The misuse of AI
The misuse of AI is not limited to just hackers put together by a country’s enemies. It can happen by the country itself as they employ AI technology for certain military tasks. These specific military tasks always include what is called collateral damage. Innocent people will be harmed.
In the use of AI to stop terrorism it is not possible to always separate the guilty from the innocent. Command decisions are made based upon the analysis of technicians who look at often unclear streaming video captured by AI drones.
This process leads to some unfortunate deaths of innocent people.
Beefing up the robot security system
There is a call by some to have the programmers of AI technology adhere to a very strong ethical code. This conclusion has been reached after a study on the dangers employing AI technology throughout life.
But it isn’t just an ethical code that is being demanded. Experts are looking to see if strict controls can be placed on AI technology. By doing this it is hoped that any of the scenarios discussed above would be avoided.
But there is a problem in implementing these strategic protections. AI is very easy to use, it is already widely in use and information is easily found. In other words, it may be too late to implement these security measures.
Secrecy & AI problems
One idea that has been proposed is that secrecy is placed over all AI research. But Jack Clark has said that action would be a very fine line to walk. It is not like the criminal use of AI has been much of a secret. Thomas Dietterich has warned companies about this facet of AI use.
Yet he wonders why top cybersecurity experts at AI companies are not involved in helping to solve the problem.
Some final thoughts
AI technology has some very good points. It can make a great contribution to life in general. But, the problem is that AI is vulnerable to the will of those who use it. AI has no ability to be ethical. It is, like guns, a tool in the hands of the person who wields it.
Controls can be circumvented, so can ethics. If the person is good, then AI will be good. If the operator is bad, then AI will be bad.