AI: The Reality

AI (Artificial Intelligence) is a hot topic as of late. With so much power available at our fingertips, it’s absolutely vital to understand the effects on ourselves and society.

Image Description: Robot Hand

Negative Effects are Dangerous

AI is an all-encompassing term for many different types of technologies. This includes things like chatbots all the way to self-driving cars. Because the term AI is so broad, it becomes  challenging to see/find the long-term data on individual types of AI. Using the AI Incident Database, Simon Mylius, a researcher at MIT FutureTech, along with others, have made a tool to enhance this database  - allowing them to attain data on different types of AI harm and classify incidents to find patterns within them. Check out this article to see the many different types of AI harm that has increased over the years.

AI overall still has its own plethora of negatives. For starters, the lack of explainable AI. For most AI, we still do not know exactly why, or how, it chooses to give certain answers. Explainable AI is the ability to understand the “why” behind the AI’s decision. Without this, we can not be sure whether or not AI is giving out biased answers or leading others to unsafe decisions. Another way AI is unsafe is its potential for social manipulation and surveillance. As shown earlier, deepfakes and other generative AI can sway people’s thoughts and opinions on politicians, celebrities, and even those who live around them. Furthermore, because of social media’s use of AI algorithms, you end up being fed the same thoughts and ideas over and over again. Often, the use of algorithms this way leads people to be trapped in a dangerous cycle of immoral ideals or practices because they feel okay when it is the only thing they are seeing repeatedly. Social media companies are not the only ones to have targeted algorithms as well. Many chatbots that people use are learning you. The chatbots take the information they know about you, and often end up creating an echo chamber of things you would like to hear. This echo chamber can lead many to what we call “AI Psychosis.”

  Social surveillance with AI stems from overuse of facial technology everywhere and other AI programs that are camouflaged as “protecting the public.” A huge one we are seeing everywhere across the US right now is Flock cameras. Flock (Safety) itself, a tech security company, leaves the liability of its devices/services in the companies that buy them. Unfortunately, this allows for the mistreatment of these cameras on a widespread scale as Flock does nothing for the prevention of this problem. Already, there have been over 20 cases of police officers using Flock to stalk ex/romantic partners and over 500,000 cases of unauthorized searches on Flock’s ALPR (automated license plate reader) data itself. Another AI program that many US police departments have experimented with is "predictive policing.” Predictive policing is an AI algorithm designed to analyze data to “help” predict and prevent potential future crimes. The overarching problem with this is bias in AI and ethical discussions on “reasonable suspicion" (which is what gives an officer the right to stop you for something). 

In addition to surveillance, lack of data privacy is still a big concern in AI. Frequently, AI systems collect personal data that is used to train the AI or help customize user experience. Currently, there is only ONE federal law in the United States that deals with AI usage (although it also applies to real images as well). There are a handful of US states that have their own individual AI laws, but overall, much of the data we put into AI systems remains unprotected on multiple levels. With all this being said, and keeping in mind the dangers of AI, it is a tool that may help in an array of fields and businesses in general.

Good Can Be Harnessed

Many of the following will have a downside - and we may have even already covered that above. This is why it’s crucial that we see the dangers of AI, like listed above, when we talk about the positives. With proper guidelines and understanding of the technology we process, we can take the good side of AI and help day to day life. 

First off, AI can do repetitive tasks making it so businesses can save time. Secondly, AI can help sort through/speed up analyzing and processing data. Additionally, because of this, AI can be beneficial in seeing trends in data thus allowing companies to make better decisions for their company. This can lead to higher customer satisfaction and better financial decisions. For some, like the Cybersecurity and Infrastructure Security Agency (CISA), AI has been used to offer safety. In CISA’s case, they use AI to detect anomalies and stop malware by constantly monitoring network traffic. Another way AI can help us is by giving accessibility where there was limited access or even none before. For many with physical disabilities, robotics (powered by AI) have aided in a plethora of things like assistive limbs, mobile wheelchairs, etc. Finally, AI has assisted in multiple ways to specific industries with many examples here. 

Overall, it’s paramount to remember that AI is meant to help guide us, not do everything for us. When it comes to any usage of it, keep an open mind and ensure you are aware of the decisions made and if it’s time to intervene. Be sure to check out our other blog, Surfing the Internet, to better inform yourself on learning things online and signs of misinformation from AI!

Now You Know!

Understanding the impacts - good or bad - is crucial in using AI and allowing it to continue developing.