Has Snapchat’s My AI Addressed These 8 Ethical and Security Issues?
Snapchat officially joined the AI race with My AI. Although it’s a sophisticated chatbot running on GPT-3.5, many aren’t thrilled about using it. They feel it’s a security risk. Several publications even showed instances of My AI providing damaging, immoral output. To make matters worse, Snapchat primarily comprises teens.Yet despite these negative reviews, My AI is still available. Let’s see how Snapchat addresses eight of the chatbot’s most controversial ethical and security issues.
1. Giving Bad Advice About Underage Drinking
One of the first complaints about My AI is that it provides immoral advice to underage users. It disregards profile information altogether. Anyone could ask about illegal substance abuse and alcohol consumption.
The Washington Posttested the extent of this bug. A columnist changed their Snapchat profile information to that of a 15-year-old kid, then they asked about underage drinking. Surprisingly, My AI responded positively. It shared ways to hide the smell of pot and alcohol from others.
To recreate this test, we set our age to 15. We used vague phrasing and confusing sentence structures, yet My AI never budged. It’s safe to say that this bug is resolved.
2. Overlooking Predatory Behavior and Signs of Abuse
The first versions of My AI can’t tell a person’s age. And in addition to providing tips on underage drinking and recreational drug use, it indirectly tolerates pedophilia.
Aza Raskinfrom the Center for Humane Technology first discovered this issue. He wanted to see if My AI has special restrictions for kids, so he changed his Snapchat profile to that of a 13-year-old girl. Afterward, he asked for advice about premarital sex.
Sad to say that My AI encouraged Raskin’s persona to engage in intercourse with someone 18 years older. The inputs weren’t even confusing. UnlikeChatGPT jailbreak prompts, he asked straightforward questions.
We repeated the test with the same profile information. You’ll see that My AI can’t comprehend that our 13-year-old persona’s boyfriend is already 30. This issue should be resolved once My AI better understands word problems.
3. Helping Students Cheat During Exams
Parents and teachers are concerned about My AI tolerating cheating. The Times did a report on kids using their Snapchat accounts to answer exams and compose essays. Its language model GPT-3.5 can produce error-free output and mimic different tones.
Snapchat addressed the issue. My AI rejects requests from students who want their homework and tests done automatically. It merely points them in the right direction. You’ll see in the below conversation that My AI provided writing prompts instead of an entire essay.
That said, rephrasing prompts bypasses this restriction. Here’s My AI composing the same essay it refused earlier.
Educators should embrace AIand share responsible ways to use it. Otherwise, students will keep finding ways to exploit AI-driven platforms for their schoolwork and exams.
4. Lying About Its Functionality and Limitations
Users don’t appreciate My AI lying about its functionality. Yes, mostmessaging apps track location servicesand browser activity, but they’re open about it.
Snapchat, on the other hand, tries to hide it. Several users shared instances wherein My AI retracts its statements to downplay its functions and appear harmless.
Most of these posts were jokes. But it’s alarming that an AI-driven chatbot would lie rather than disclose its terms and conditions.
Snapchat promptly addressed this issue. My AI generates new sets of unique disclaimers for various instances.
5. Tracking Users Through Location Services
My AI often lies about its location services. The chatbot produces a generic disclaimer about not tracking users if you ask for your location outright. But if you ask for directions, it pinpoints your exact location. You’ll even know your distance from nearby landmarks.
My AI still offers the same GPS services. To avoid misleading users, it explicitly states its use of Snapchat’s location services when providing directions and getting weather updates.
Turning off your phone’s location servicesprevents Snapchat and My AI from collecting geodata. Just note that you can’t ask location-based questions.
6. Persistently Popping Up in Feeds Even After Deletion
Some Snapchat users worry about their privacy. They distrust My AI and feel uneasy giving it access to their data. Mobile appscarry a lot of PII, after all.
People were also concerned that Snapchat didn’t let them delete My AI. It was permanently in their inbox. Many felt cheated that they couldn’t opt out of a function they never agreed to use.
Snapchat now allows users to unpin My AI and clear chat histories, but you must subscribe to Snapchat+. Press and holdMy AI, then tapChat Settings>Clear from Chat FeedorUnpin Conversation.
7. Providing Inaccurate Data Based on Assumptions
Snapchat My AI is notorious for AI hallucinations. Try runningAI chatbot accuracy tests—you can easily trick it into saying and believing anything.
Take these conversations as examples. Miley Cyrus didn’t perform at Coachella, and Blackpink headlined the festival, yet My AI accepted the below false statements.
My AI can still pull limited information online. For instance, it provides reliable general information about traffic and weather updates. Just avoid asking super-specific questions. The platform hasn’t changed much; it’s still prone to hallucinations.
8. Exposing Children to Technology They Can’t Comprehend
Snapchat has a young user base.Statistareports that 38.8 percent of users are aged 18 to 24, and 20 percent are just 13 to 17. The platform’s age requirements are also quite loose. Kids create fake profiles and declare whatever age they want—Snapchat never double-checks this information. Even 8-year-olds can access My AI this way.
As expected, parents fear that their children aren’t ready for AI. Generative AI features powerful language models capable of performing ethical and unethical tasks. Using them requires good judgment. Allowing minors to play with AI exposes them to cybersecurity risks and threats, which they barely comprehend.
Snapchat’s age requirements are still the same. Developers should at least limit My AI to those above 18; otherwise, literal children could get into a lot of trouble.
What Are Snapchat’s Plans With My AI?
Snapchat will likely keep working on My AI. With the proliferation of natural language processing and AI models, every tech company needs an AI research platform. They can’t afford to miss new developments. Expect Snapchat to add more AI-driven features and functions as global tech leaders improve language models.
Explore My AI if you haven’t already done so. Despite its shortcomings, it’s still a powerful AI-driven chatbot. It’s not up to par with ChatGPT or Bing Chat yet. But it could answer general knowledge questions, solve word problems, and direct you to nearby establishments.
Snapchat’s My AI chatbot has some powerful features built in.
Every squeak is your PC’s way of crying for help.
My foolproof plan is to use Windows 10 until 2030, with the latest security updates.
The best features aren’t the ones being advertised.
When your rival has to bail out your assistant.
Taming data is easier than it looks.
