Key takeaways
- You can make an AI app as a beginner by calling a model API, and train a small model yourself with Python and free tools.
- Start with one useful project, such as a question-answering bot over your notes, rather than training a model from scratch.
- AI images are easy to make free in Gemini, ChatGPT and Copilot; AI video is limited on free tiers.
- ChatGPT's video generation comes from Sora; Gemini's from Veo; Adobe's from Firefly in Express.
- Label AI-generated media and never create misleading images or videos of real people.
Yes, you can make AI. As a beginner, the fastest route is to build an AI application on top of an existing model through an API, such as a chatbot that answers from your college notes; with some Python, you can also train small machine learning models yourself. You can also make AI images and videos with free tools. This page covers building your first AI, why learning AI and ML is worth it, and how to make AI photos and videos with ChatGPT, Gemini and Adobe Express, including what is free.
Can you make your own AI?
Yes. You can build AI applications by connecting to models such as Gemini, GPT or Claude through their APIs, and you can train your own smaller models with Python libraries such as scikit-learn and PyTorch. Training a large model like ChatGPT from scratch is beyond individuals.
Three levels of "making AI":
- Use a model through an API: build a chatbot, summariser or document assistant. Hours to days.
- Train a small model: a spam classifier or image recogniser on your own data. Days.
- Fine-tune an open model: adapt a model to your task with LoRA on a single GPU. Days to weeks.
Most real products are the first kind, done well. That is how I built RAG.NextUpgrad and DocuLens AI: existing models, with retrieval, validation and careful engineering around them.
How do I make an AI as a beginner?
Learn basic Python, get a free API key from Google AI Studio, and build one small, useful AI app: for example, a study assistant that answers questions from your notes and cites the page.
A beginner path over a few weeks:
- Week 1: Python basics, variables, functions, lists and dictionaries.
- Week 2: call an LLM API and print a response; then add a simple web interface with Streamlit.
- Week 3: add your documents, split them into chunks and retrieve the relevant ones for each question.
- Week 4: deploy it for free and share it with friends for feedback.
Alongside, do a short machine learning course to understand training and evaluation. Build first, then fill gaps in theory as you need it.
Why learn AI and ML?
Because AI and machine learning are becoming part of almost every software product and many non-technical jobs, and people who can build with AI are in demand. Learning them also teaches durable skills: data handling, statistics, programming and careful evaluation.
Also asked as: why ai and ml
For students, AI skills open roles such as AI engineer, ML engineer, data scientist and AI product developer, and make every other engineering role more productive. Even outside tech, knowing how models work helps you use them well and spot their mistakes. The field rewards builders: a student who ships one real AI project learns more than one who completes many courses. Start with a problem you care about, and let it pull you through the learning.
Can you make AI images?
Yes. Gemini, ChatGPT and Microsoft Copilot can all generate images from a text description for free within daily limits, and dedicated tools such as Adobe Firefly and Canva's AI features do the same.
Also asked as: how to make an ai for beginners
Write prompts with subject, setting, style, lighting and composition, for example "a flat illustration of students at a hackathon, laptops and sticky notes, bright colours, top-down view". Ask for changes in follow-up messages instead of starting over. For text inside images, such as posters, newer models do much better than older ones but still make mistakes, so check spelling. Respect copyright and trademarks: avoid generating logos or characters you do not own, and label AI images when they could mislead.
Can you make AI photos?
Yes. You can generate realistic photos from text, and edit your own photos with AI: change the background, adjust lighting, extend the frame, remove objects or apply a style. Gemini, ChatGPT, Adobe and many phone apps offer these features.
For edits, upload the photo and describe the change precisely, such as "replace the background with a plain light grey studio wall, keep the person unchanged". Faces can shift slightly in AI edits, so check carefully before using images for anything official; never use AI edits for ID or passport photos. Never generate or edit photos of real people in ways they have not agreed to. Many tools add visible or invisible watermarks to mark AI images, which is a good thing for trust.
Can you make AI videos?
Yes. Text-to-video and image-to-video models, such as OpenAI's Sora, Google's Veo, Adobe Firefly and others, generate short clips from a prompt or a still image. Longer videos are built by editing several clips together.
Generated clips are short, so plan a video as a sequence of shots. Describe each shot like a director: subject, action, setting, camera movement, lighting, style and sound. Expect to generate several takes and choose the best. For explainer or tutorial videos, real screen recordings and your own voice usually work better than generated footage, with AI clips as B-roll. Label AI-generated video when you publish it, and check your platform's rules on disclosure.
Can you make AI videos for free?
Partly. Free options exist but come with tight limits, watermarks or short clip lengths. Some apps offer a few free generations a day, and video editors such as Clipchamp, CapCut and Adobe Express include free AI features.
Ways to make AI-assisted videos without paying:
- Write the script with a free chatbot and record your own voice and screen.
- Generate images for scenes for free, then animate them with simple pans and zooms in a free editor.
- Use free daily credits in video generation tools for a few key shots.
- Use free text-to-speech for narration, labelled as synthetic where relevant.
Free offers change frequently, so check each tool's current plan page.
Can you make AI videos with ChatGPT?
ChatGPT itself writes, plans and creates images, while video generation comes from OpenAI's Sora, available through its own app and website for eligible users and regions. ChatGPT can still produce scripts, storyboards and prompts for Sora.
A workflow: ask ChatGPT for a 30-second script and a shot list, then turn each shot into a detailed Sora prompt, generate the clips, and edit them together with narration and captions in a video editor. ChatGPT can also write ffmpeg commands to assemble images into a slideshow video. Availability and limits for Sora depend on your plan and country and have changed since launch, so check OpenAI's current information.
Can you make AI videos on Gemini for free?
Video generation in Gemini, powered by Google's Veo models, has mainly been part of paid Google AI plans with daily limits, though Google has run trials and student offers. Free accounts can still use Gemini to plan videos and generate images.
If you have access, choose the video option in Gemini, write a detailed scene description and generate a short clip with sound. Without access, use Gemini for scripts, shot lists and still images, then assemble a video in a free editor. Google has offered free plan periods to eligible students in some countries, including India, so check the official Gemini pages for current offers. Generated videos carry Google's SynthID watermark.
Can you make AI videos with Adobe Express?
Yes. Adobe Express includes AI features powered by Adobe Firefly, including generating short video clips from text prompts, plus templates, animation, captions and editing tools to turn them into finished social videos.
Adobe Express is beginner-friendly: pick a template, generate or add clips, add text and music, and export for Instagram, YouTube Shorts or presentations. Free accounts get limited generative credits, and paid plans include more. Adobe trains Firefly on licensed and public-domain content and positions it as designed for commercial use, which matters for business projects; still, read the current terms. As with any AI video, review the output for errors and label it where appropriate.
Sources
GenAI engineer and AI product builder with 2+ years shipping production-grade AI systems: RAG pipelines, fine-tuned LLMs, hybrid retrieval and multi-modal apps across vision, speech and OCR, architected end to end from ingestion to deployment. Leads engineering for SCULT INDIA's 14-member team, founded the 500+ member TechVerse Enclave community and has mentored 200+ students. Three hackathon first prizes: Changethon 2025 (IIT Roorkee), Product Genesis at Vividhotsava 2025 (CSJMU) and BYTEBATTLE (MeetKats).
Open to GenAI roles, hackathon judging, mentorship sessions and guest talks at colleges. For campus sessions the inviting institution covers the visit end to end: travel (a First AC train ticket for journeys up to five hours, a flight for anything longer), accommodation, meals, and pickup and drop; the honorarium and workshop fees are negotiable on WhatsApp or on the call, and requests are welcome from institutions that can cover the visit this way. Email pranjulrathour41@gmail.com.
