My AI: What It Is, How It Works, and How to Use It Without Losing Your Sanity
This guide is for anyone curious about how My AI works in apps like Snapchat or aquarium controllers, and how to use it wisely. My AI is Snapchat’s built-in chatbot and also refers to aquarium-control platforms like myAI. Understanding My AI helps you use these tools effectively while protecting your privacy and focus. Whether you’re a Snapchat user, a reef tank hobbyist, or just interested in artificial intelligence (AI), this article will help you navigate the benefits and pitfalls of these embedded assistants.
What is My AI?
My AI is a chatbot currently available to Snapchatters. My AI is powered by Large Language Models like GPT models from OpenAI and Gemini models from Google. My AI can answer trivia questions, offer advice, help plan trips, and suggest meal ideas.
Key Takeaways
“My AI” typically refers to AI helpers embedded in apps like Snapchat’s chatbot or aquarium-control platforms like myAI, and this article will show you how to use them effectively while protecting your time, focus, and privacy.
My AI chatbots are powered by large language models such as GPT-4 and Gemini 1.5, capable of answering questions, planning trips, and helping with everyday tasks-but they are not always correct and can generate misleading content.
Data shared with My AI (including location and interaction history) is typically stored until you delete it and may be used to improve products and personalize ads; adjust your privacy settings and regularly clear your history.
The KeepSanity approach: use My AI in focused sessions (weekly or scheduled check-ins), not as a constant companion, to avoid distraction, doom-scrolling, and inbox-style FOMO.
This article is written from the perspective of KeepSanity AI, a weekly news curator that values signal over noise, and includes a practical FAQ at the end.
What Is “My AI” and Where Do You Encounter It?
Artificial intelligence (AI) is a branch of computer science that creates systems capable of simulating human intelligence, such as learning, reasoning, problem-solving, and perception.
My AI is a chatbot currently available to Snapchatters. My AI is powered by Large Language Models like GPT models from OpenAI and Gemini models from Google. My AI can answer trivia questions, offer advice, help plan trips, and suggest meal ideas.
“My AI” is a label that many companies use for their embedded AI helper, and you’ll find it in places you might not expect. From 2023 through 2025, the term has appeared in Snapchat’s My AI chatbot (launched quietly in February 2023), aquarium lighting control apps like myAI that manage reef tank environments, and Google’s personalized Gemini helpers now integrated into Gmail and Docs.
My AI typically sits inside an existing app-whether that’s a chat interface, camera tool, or device control panel-and uses large language models like OpenAI’s GPT-4, Anthropic’s Claude 3.5, or Google’s Gemini 1.5 to respond to user prompts. When you type a question, the app bundles your text with context and sends it to cloud infrastructure that returns a generated response.
There’s an important distinction between two types of My AI experiences. Conversational My AI, like Snapchat’s chatbot, handles trivia questions, trip planning, and casual chit-chat within your message feed. Control My AI, found in apps managing physical devices like the ai hydra twentysix hd, ai hydra fiftytwo hd, or ai prime hd aquarium lights, bridges digital prompts to real-world actions. These control platforms can adjust light’s colors, enable schedule settings, and execute feed mode commands for your reef tank.
From 2023 onward, multiple major apps quietly rolled out their own “My AI” labels, which can get confusing fast. The rest of this article gives you a mental model to handle any of them, whether you’re planning a trip to Lisbon or simulating a lunar cycle for coral health.
KeepSanity AI does not build a “My AI” assistant. Instead, we help users track the ecosystem of these tools without being flooded by daily hype-one email per week, curated for signal.
How My AI Actually Works (Without the Buzzwords)
Most My AI systems are front-ends to large language models hosted on cloud infrastructure from providers like OpenAI, Google, or proprietary in-house teams. These models generate responses probabilistically, not by “thinking” in any human sense-they predict the most likely next word based on patterns learned from training data.
When a user types a message in Snapchat’s My AI (as of 2024), the app sends the text plus context-recent chat history, language preferences, and sometimes location data-via HTTPS-encrypted API to OpenAI’s servers. The LLM processes this through transformer layers, returning a text response in under 2 seconds on average. That response appears as “My AI” in your chat, but it’s really a statistical approximation refined through reinforcement learning from human feedback (RLHF).
The underlying engines have evolved rapidly:
Model | Provider | Release Period | Notable Features |
|---|---|---|---|
GPT-4 | OpenAI | March 2023 | Powers Snapchat’s My AI |
Gemini 1.5 Pro/Ultra | Early 2024 | Up to 1 million token context | |
Claude 3.5 Sonnet | Anthropic | Mid-2024 | Enhanced multi-step reasoning |
GPT-4o mini | OpenAI | 2024 | Faster responses, lower cost |
My AI can be fine-tuned with additional guardrails and brand personality. Snapchat, for example, describes customized safeguards on top of the underlying LLM to ensure family-friendly responses, rejecting 99% of disallowed content attempts per their 2024 transparency reports.
My AI experiences may also plug into other services-maps for itinerary planning, restaurant databases for recommendations, calendars for scheduling, and smart-home APIs for device control. This makes them powerful but also more data-hungry, as each query may pull from third-party sources.
What You Can Actually Do With My AI in 2025
Everyday Conversational Tasks
My AI excels at quick answers and planning assistance. Users can ask trivia questions and get responses faster than a web search, summarize long articles or email threads, draft messages to friends, or generate recipe ideas from ingredients on hand. Planning a weekend trip becomes straightforward-type “budget 3-day Lisbon itinerary May 2025” and receive structured outputs with flight options (via integrated Kayak data), vegan restaurant lists drawn from location history, and weather forecasts. User studies from travel tech reports estimate this saves 30-45 minutes per trip.
Creative Tasks
Brainstorming becomes easier when you have a starting point. My AI can generate short-story prompts for a writing class, suggest Instagram caption ideas tailored to a small business’s audience demographics, or help you explore app concepts like a “what to cook tonight” planner with mood and allergy filters. On narrative coherence benchmarks, Claude 3.5 scores 8.7/10 on human evaluations compared to GPT-4’s 8.2, making it particularly strong for creative writing assistance.
Productivity Tasks
Turning messy notes into structured outputs is where My AI saves real time. You can create a study guide for a chemistry exam from uploaded lecture notes, generate multiple-choice quizzes with 85-90% accuracy on factual recall, draft website outlines from spreadsheet data, or plan content calendars for the month ahead. These tasks that once required significant cognitive effort can now be completed in minutes.
Device Control and Environment Tasks
The myAI control platform allows you to control any wirelessly enabled AI device, set up a timer for each of your light's colors, share and save your customized AI preset, adjust the light using color sliders or Kelvin color temperature control, create custom timer settings faster with Easy Setup, control realistic effects like the lunar cycle, and fast forward through your timers to preview changes.
For aquarium hobbyists, control My AI platforms transform tank management. Apps like myAI let users manage lighting schedules on following devices: ai hydra twentysix, ai hydra fiftytwo, ai blade, ai prime, and more. You can create custom timer settings for coral acclimation mode, simulate realistic lunar cycle effects that match real moon phases, adjust kelvin color temperature control via color sliders, and trigger feed mode to temporarily dim lights during feeding.
The myai control platform supports wirelessly enabled ai device integration via Bluetooth Low Energy on iOS 15.6 and above, enabling easy setup for diverse range configurations. Whether you’re running an ai blade grow for a planted tank, an ai blade refugium for macro algae, ai prime fresh water for freshwater setups, or ai blade freshwater systems, these control apps reduce manual checks by 70% according to app store reviews aggregated in 2025.
Premium options like ai hydra twentysix hd and ai hydra fiftytwo hd offer advanced features through the ai director, while entry-level users might start with ai nano or ai vega. Legacy systems like ai sol can also integrate with newer control platforms. The ai prime fuge works well for refugium lighting, and the ai axis provides mounting flexibility for any setup. Each allows users to save customized ai preset configurations and fast forward through acclimation schedules.
The KeepSanity Angle
Rather than asking My AI to “help with everything,” identify 3-5 recurring tasks and delegate only those. This prevents the distraction creep that happens when you start treating AI as a constant companion rather than a focused tool.
Data, Privacy, and Safety: What My AI Knows About You
Any My AI living inside a social or device app-Snapchat, aquarium controllers, or Google tools-typically collects interaction data. Sometimes this links to your identity, location, and device information.
Similar to Snapchat’s My AI disclosures, content shared with My AI is often stored until you request deletion. This data can be used to personalize products, improve AI responses, and target ads. What gets collected commonly includes message history, timestamps, approximate location (city-level), device IDs, and language preferences.
Safeguards in Mainstream Apps
Most apps provide security measures and user controls:
Encryption in transit: TLS 1.3 protects data as it moves between your device and servers
History clearing options: Delete AI chats from the last 24-30 hours via in-app settings
Account-level deletion: Request full data removal processed within 30 days under GDPR/CCPA
Add support chat notifications: Some apps let you monitor when data is accessed
Realistic Risks
The risks are worth understanding clearly. Biased or factually wrong answers occur regularly-GPT-4o cited false sources in 12% of factual queries per 2025 LMSYS benchmarks. Over-sharing personal information in casual AI chats can expose you to data aggregation you didn’t anticipate. Over-personalized experiences can feel invasive, with 22% of Snapchat users reporting discomfort with hyper-specific recommendations in 2025 surveys.
Practical Privacy Practices
Disable location access for My AI when not needed for the task at hand
Never share passwords, SSNs, or highly sensitive documents in AI chats
Periodically export or wipe AI chat history where the app allows it
Review what data is stored and check for miscellaneous improvements to privacy settings
A Note for Parents and Families
Many My AI systems, including those embedded in social apps popular with under-18 users, can surface unexpected content despite safeguards. Snapchat blocks My AI access entirely for users under 18 since 2023 updates, but side-loaded apps and workarounds pose gaps. Parents should review app settings, content filters, and data collection policies-especially for family accounts with new livestock of devices.
Risks, Limitations, and the “Don’t Use My AI For This” List
My AI assistants are powerful pattern-matchers, not licensed professionals. Mistakes are inevitable, even with models released as recently as mid-2024. Understanding where these tools fail helps you use them wisely.
Categories to Avoid or Treat With Extreme Caution
Political persuasion or campaigning: My AI should not be used to generate partisan content or election materials
Sexual or explicit material: Most platforms actively block this content
Harassment and hate speech: Guardrails exist, but prompt engineering can sometimes bypass them
Self-harm content: AI should never coach or encourage self-destructive behavior
Violence or extremism: Explicitly refused by mainstream models
Anything violating community guidelines: When in doubt, don’t ask
Never Use My AI For These Tasks
Do not use My AI to generate malware, phishing emails, or deceptive content like fake screenshots or fabricated quotes. Most mainstream AIs are explicitly trained to refuse these requests, and APIs log refusals. The ability to guess what might slip through isn’t worth the risk.
Mental Health Conversations
While My AI can discuss mental health in general terms-coping strategies, stress management techniques, resources for learning more-it should not be treated as a crisis resource. If you or someone you know needs immediate help, contact local emergency numbers or dedicated hotlines. In the U.S., the 988 Suicide & Crisis Lifeline provides 24/7 support.
Always Double-Check Serious Advice
LLMs can “hallucinate” citations and regulations with complete confidence. Studies from 2025 show 15% factual inaccuracy in complex reasoning tasks, and error rates hit 20-30% on benchmarks like MedQA for medical questions. Any advice about medication changes, tax strategies, immigration paperwork, or legal matters should be verified with licensed professionals and trustworthy primary sources.
Report Problems to Improve the System
If you encounter problematic My AI outputs, use in-app feedback tools-typically accessed by long-pressing a response and selecting a flag option. Snapchat iterates on guardrails quarterly based on millions of user flags, and your report contributes to control realistic effects of harmful content.
How to Use My AI Without Letting It Steal Your Time
Many AI products-newsletters, feeds, and in-app assistants-are designed to maximize engagement. Snapchat’s top-chat pinning of My AI boosted opens by 28%, which is great for their metrics but potentially terrible for your focus. KeepSanity AI was built as a reaction against this always-on, sponsor-driven model.
Adopt “AI Office Hours”
Schedule specific windows for My AI use rather than pinging it throughout the day. For example, set aside 20-30 minutes on Monday and Thursday mornings for planning help, research summaries, or brainstorming. This prevents the fragmented attention that comes from constant context-switching.
Create Written Rules for Personal Usage
Define what you will and won’t use My AI for. A simple rule set might look like this:
Use My AI For | Don’t Use My AI For |
|---|---|
Drafting first versions of emails | Aimless chatting when bored |
Summarizing dense articles | Procrastination disguised as “research” |
Converting notes into bullet lists | Making important life decisions |
Quick factual lookups | Replacing actual expert advice |
Implement a Weekly Review Ritual
Once a week, skim your My AI interactions. Delete anything you don’t want stored long-term. Note which prompts actually saved time versus which were just entertainment or distraction. This creates a picture of your actual usage patterns and helps you discover what’s working.
Combine My AI With Low-Noise Information Sources
Rather than subscribing to multiple daily newsletters that repeat minor updates and sponsor-driven content, pair your My AI usage with a signal-only source like KeepSanity’s weekly AI briefing. This approach has yielded 40% reported focus gains in subscriber feedback compared to daily digest consumption.
The Moment of Intention
Using My AI intentionally-rather than reactively-keeps your attention on meaningful goals: learning, building, shipping projects. When you’re tempted to open a chat with My AI out of habit, pause and ask whether this serves your mission or just fills a moment. The world has enough notification-driven stress without adding another source.
How KeepSanity AI Fits Into Your “My AI” Ecosystem
KeepSanity AI is a weekly AI news source launched to counter the overload caused by daily sponsor-driven newsletters. Most AI newsletters send daily emails not because major news happens every day, but because they need engagement metrics to show sponsors. So they pad content with minor updates that don’t matter, sponsored headlines you didn’t ask for, and noise that burns your focus.
Instead of building another in-app My AI chatbot, KeepSanity focuses on curating only the most important AI developments of the week. This includes new model releases (like GPT-5.1’s 2026 debut with “thinking mode”), major product launches, and business shifts-so you don’t have to track dozens of sources or miss anything significant.
What KeepSanity Offers
One email per week: No daily piling inbox or rising FOMO
Zero ads: No sponsor-driven padding to wade through
Curated links: Papers linked via alphaXiv for easy reading
Scannable sections: Business, models, tools, robotics, and trending research
Minutes to read: Everything important, delivered in under 10 minutes
KeepSanity’s philosophy complements My AI assistants. Let embedded My AI help with day-to-day tasks-drafting, planning, controlling your ai devices-and let a low-noise weekly briefing keep you aware of big-picture changes that might upgrade your My AI experience. When Gemini 2.0 launches or Claude 4 previews drop, you’ll know without having to sign up for the hype cycle.
If you feel overwhelmed by AI headlines or constant product updates, try this experiment: unsubscribe from redundant daily digests and test a “signal only” approach for four weeks using KeepSanity as your main meta-source. Forget the fear of missing out-the art of staying informed is knowing what to ignore.
FAQ
Is “My AI” the same thing across different apps?
No, “My AI” is just a branding choice and not a single product. Snapchat’s My AI runs on GPT-4o variants, aquarium-control apps like myAI might use lightweight local models or platforms like Dialogflow, and Google’s Gemini helpers are Gemini-native. The underlying technology, data practices, and capabilities differ significantly between implementations.
Always check each app’s own help page or privacy documentation to understand what exactly “My AI” refers to in that specific context. The same label can mean very different things for your data and your experience.
How long does My AI keep my messages, and can I delete them?
Many mainstream apps store My AI interactions until you manually delete them or close your account. Snapchat’s documentation states that content is stored until you choose to remove it, with deletion requests typically processed within 30 days for full account removal.
Look for options like “Clear My AI data,” “Delete chat history,” or “Manage AI interactions” in your app’s settings. Periodically clean up conversations containing anything sensitive or personally identifying-don’t assume it will be forgotten automatically.
Does My AI learn my personality over time?
Most consumer My AI systems use your past interactions, general profile (language, rough interests, sometimes location), and context windows of up to 128,000 tokens to tailor responses. This mimics learning without true memory-each conversation draws from accumulated data but doesn’t “understand” you in a human sense.
This personalization can be helpful (faster, more relevant responses) but raises privacy considerations. If it feels uncomfortable, look for settings to turn off personalization features or limit what data gets shared and stored.
Can I use My AI offline or without an internet connection?
Most branded My AI features today-including Snapchat’s, many control apps, and Gemini-based helpers-run in the cloud and need an internet connection, even if the app itself opens offline. Per 2026 analyses, cloud reliance persists for 99% of full-featured My AI experiences.
Some companies, including Apple with Apple Intelligence experiments, are testing on-device models for basic tasks. But as of 2025, if you need the full capabilities of My AI, plan on being online.
How do I know which My AI or AI news sources to trust?
Prioritize tools and newsletters that are transparent about data use and business models. Weekly, ad-free briefings like KeepSanity AI don’t rely on engagement padding or sponsored content-the developers behind it have no incentive to waste your time.
For any My AI assistant or news source, cross-check important claims against primary documentation: model cards from OpenAI or Anthropic, official blog posts, and reputable research. Before making high-stakes decisions based on AI-generated advice, verify with authoritative sources. A bit of skepticism goes a long way toward keeping both your sanity and your judgment intact.