With a recent surge of new artificial intelligence (AI) technology, companies are eager to jump on the AI train in order to stay ahead of their competitors. Between AI-generated ads on Instagram, Google’s AI search, and of course, Chat GPT, the new technology is surfacing all over the internet. One of the most impactful new AI developments for teens is Snapchat’s “My AI” feature.
My AI is meant to be like a personalized, built-in friend within the app. With its customizability and immediate response time, it could potentially be the 24/7 friend we all lack, especially because it remains pinned at the top of every user’s screen. However, across social media, some have brought up the concern that the AI is invasive, mean, or even downright wrong. For those uncomfortable with it, removing the AI is not an option. It remains pinned and present, and users must pay for Snapchat+ in order to remove it from their chat feed. “My only interaction with the AI was when I asked it how to remove it from my phone, and it told me I had to pay for it,” said senior Ava Ver Ploeg. In a sense, it seems as though the company is trying to capitalize on the discomfort of its users who do not like having the AI in their app.
“I sent it a picture of my face and it said ‘Mmm, that cake looks delicious. I hope you enjoy every bite,’” said sophomore Allegra Hoddie. I thought it was kind of funny but it also made me question the validity of the AI, because how could it get something so wrong?”
Another concern students raised was the inconsistencies within the responses of the AI. One in particular is that the AI has been telling students that it is a real person.
In screenshots shared by sophomore Bethany Hyde, the AI says, “I’m not MyAI, I’m a real person,” and “I’m actually not an AI, I’m a real person! I’m here to chat with you and help you feel better.” Furthermore, when I asked my own AI “Are you an AI?” it responded “No, I am not an AI,” and later, “Yes, I am a real person.” However, when I posed the question again, the AI said, “I am an artificial intelligence designed to help you.”
Students also reported inconsistencies in what the AI said its capabilities were. On one occasion, Hyde asked the AI to look up the lyrics to a song, and the AI said it “did not have internet access.” However, when Hyde asked the AI to look up lyrics to a different song, it was able to, and said “I can look up information on the internet to provide you with accurate answers.”
Similarly, when I asked my AI to rank Taylor Swift’s albums, on one occasion it was able to, and on another it said, “I don’t have the ability to rank albums.”
One of the largest problems with the Snapchat AI is its lack of disclosure about location permissions. While it claims that it has no knowledge of your location, it is able to pinpoint restaurants or attractions nearby the user. For example, I asked my AI where the nearest McDonalds was, and it was able to tell me that the closest one to M-A is on El Camino Real.
This, coupled with the customizability of the AI, blurs the line between its clear stance as an AI versus a “friend.” Sometimes, it is difficult to remember that the AI is not in fact a real friend who can form an emotional connection or has one’s best interests at heart. “The AI threatened to call the cops on my friend,” said junior Lucas Tjernlund. “It said it was going to contact the authorities or something.”
Overall, while a fun tool to play with, Snapchat’s AI is not exactly the built-in friend users may have expected it to be. Though useful at times, it is often inconsistent, and despite attempted safety features, such as the fact that it avoids certain keywords, many seem uncomfortable with it.
