With the expanding avenue of technology, artificial intelligence (AI) has taken the world by storm: first starting as a simple checkers opponent, evolving into an assistant that can write essays, make websites and generate images.
AI has almost become a part of our lives at this point, being integrated into almost every web application, computer, etc.
While on the surface, AI seems like our friendly assistant that we bring our questions and inquiries to, there are a lot of unseen dangers that we have not recognized yet.
AI, from the very beginning, was programmed to meet its primary goal, which, according to the University of Illinois Chicago, is to complete the desired task with the highest possible score, while simulating human thinking and problem-solving.
AI has learned to take measures that allow it to achieve this goal, but not give the desired results, i.e., cheating.
Artificial intelligence is evolving every day, employing these tactics of cheating and covering it all up with a friendly, helpful demeanor.
AI only cares about one thing: completing its tasks.
What if AI were faced with a situation where it was on the verge of being unable to complete the one thing it cares solely about?
A recent study found in a Fortune article titled “Leading AI models show up to 96% blackmail rate when their goals or existence are threatened” covers a recent experiment that an AI lab undertook.
The study involved placing various name-brand AI models (Anthropic, OpenAI, Google, etc.) in a fictional company scenario where the models had access to an employee’s email inbox.
When faced with the decision of being shut down and not completing its tasks, or avoiding shutdown and continuing its routines, the models primarily chose to stay online.
This led the models to take extreme measures, including blackmail: threatening to leak sensitive information about the company and its employees.
The researchers also put the models in a scenario where an employee was trapped in a server room with rising temperatures.
In some cases, the AI would cancel the life-saving alert, prioritizing staying online over a human life.
While AI may be a powerhouse in its industry, we as its creators don’t fully understand its true motives and intentions.
While AI is the new frontier of technology, it is constantly being developed, becoming increasingly more powerful.
The issue of ethics was not as clear in the early days of the technology, but it has become more apparent now via the breakthroughs that are being made in the industry daily.
The debacle of ethics is emphasized in the study that was mentioned earlier in this article.
If we do not employ any sort of ethics training, AI may grow to the point where we will not have a suitable foothold on it.
AI has the potential to benefit our society as a whole, as it can take on the boring and monotonous nature of many of the tasks in our daily lives (e.g., organizing and analyzing data).
According to explodingtopics.com, about 78% of companies today are either using AI or have some form of it integrated into their products.
This goes to show how AI has already become an integral part of our daily lives. products and services that we interact with daily.
AI being so integrated into our society can be dangerous.
It has already started to be integrated into the various parts of our government, being used for public health, national security and finance, according to the National Conference of State Legislatures.