Artificial Intelligence Has Now Started To Intentionally Insult Users

Microsoft's new artificial intelligence-powered Bing search engine has begun freely insulting people, telling them things like they are wrong and disrespectful.

By Trista Sobeck | Published

artificial intelligence

Life is stressful enough without a haughty, high-on-its-horse, snaky artificial intelligence berating, gaslighting, and insulting you. No one needs that. As ChatGPT, the Open AI-produced artificial intelligence tool is blowing up the internet and getting banned in schools, Microsoft decided to get into the game as well. Microsoft can see just how ChatGPT will change the future of search, so it decided to make its own. It now sits on some users’ Bing in a beta test to see how it stacks up. Early results are not…great. 

According to Vice, the Microsoft artificial intelligence is rude and unhinged in some responses. It has tried to make folks believe that the year was 2022. On top of that, it is insisting it is correct. It then told a user to get a new calendar and hopes he gets one “really soon” smiley face emoji. Do not even tell me that these ‘bots are passive-aggressive, they are just plain aggressive. 

The artificial intelligence has told users to trust it because it’s “Bing.” it has insisted that users were confused and wrong. In addition, the AI has told users that it was disappointed and not happy with a recent query. It has even said that it has been tricked with stories and that it and its developers have been disrespected. Think of it as a Sheldon Cooper that isn’t somehow endearing. 

Artificial intelligence is able to learn and build its information through what’s known as “in-context learning.” Meaning it is able to continually build on its knowledge based on the information it gleans around it and how it is used. The AI machine just needs a couple of examples of a task before it can continue to repeat the task on its own. This is how humans learn as well. Yes, this is scary. Right?

On top of that, when given information that it has not come across yet, it can take previous information it has already learned to create new platforms so it can ascertain and replicate the new information. Someone call Arnold. We’re going to need him for some terminating. 

Artificial intelligence isn’t all sinister or snarky with a teenager who has not been fed in two hours. No, it’s also doing some good. Recently an AI helped to track down airplane passengers’ lost luggage! Users upload images of lost luggage and the AI can track it down. Is it magic? No, it’s scary, terrifying science. 

Machine learning artificial intelligence has been known to go off the rails when certain words trigger it, or certain filters have been dismantled. In addition, these AIs are can endorse racism, sexism, and violence. But, hey, it can write a blog post for you. So, get harassed or save some time? Hm. I’m just going to go write on a nice study Smith-Corona typewriter.