Choosing, licensing and treating images

0:00 Choosing and licensing images

Here's a quick question. What's the first thing you notice in most social posts and ads? Not the headline. The image. A strong image carries the whole message. A weak one, generic, low quality or unrelated, undermines even perfect typography. And there's a second question that matters just as much: are you actually allowed to use it? Because found it on Google is not a license. In this lecture, you'll learn what makes an image work, where to get images and what their licenses usually allow, how AI images and provenance fit in, how to treat images consistently, and a scorecard you can use to choose. By the end, you'll pick images that work and that you can defend.

0:52 Why it matters

Why does image choice matter so much? Because images do the heavy lifting in a fast feed. They stop the scroll, set the emotion, and signal whether the brand is real. And using images legally is a professional responsibility. Copyright usually belongs to the creator from the moment an image is made, whether or not there's a notice on it. Think of images like music in a café. The café can't just play any song it likes in public without the right license. Images in marketing work the same way. A great image you can't legally use isn't a great image. It's a future takedown notice.

1:38 What makes an image work

Let's start with what makes an image work. Use five questions. Relevance: does it show the actual product, person, place or outcome, or a vague metaphor? Clarity: is there one clear subject that reads at thumbnail size? Emotion: does it create curiosity, desire, empathy or trust? Authenticity: does it look real and specific, not like the over-lit handshake or the perfect team laughing at a laptop? And space: is there room for text without covering the subject? Here's the key idea. Original photos of your product, workspace, face and real customers, with permission, usually beat stock, because they're specific and credible.

2:22 Sources and licenses

Now sources and licenses. This isn't legal advice, and laws differ between countries, but here are the patterns. Your own photos: you generally own them, but you need permission from identifiable people. Free stock libraries: often permissive, but with restrictions like no selling unaltered copies or implying endorsement. Paid stock: standard or extended licenses, with limits on things like print runs, merchandise or templates for resale. Creative Commons: various types, some requiring attribution, some banning commercial use or edits. AI-generated images: governed by the tool's terms and local law. And images found online: copyrighted by default. For commercial use, identifiable people usually need model releases, and some property needs property releases.

3:10 AI images and provenance

Let's talk about AI images and provenance, because they're now part of every image decision. AI-generated images are governed by the tool's terms and the law where you work. In the US, purely AI-generated material generally isn't protected by copyright without sufficient human authorship, so a competitor might be able to reuse it. Many stock libraries now sell or label AI-generated images with their own rules, so read the license for each one. And many cameras, editors and AI tools can attach Content Credentials, provenance data based on the C2PA standard. Keep it when you export, and never strip credentials from images you license. Most importantly, never present AI images as real places, products or people.

4:00 Worked example 1: a coach's hero image

Worked example one, simple. A life coach uses a stock photo on her website hero: three people laughing at a laptop. Let's score it. Relevance: zero, it's not her, her clients or her work. Clarity: one. Emotion: one, generic happiness. Authenticity: zero. Space: one. Four out of ten. And the license? She downloaded it from a free site and can't find the source. Now the replacement: a photo from one of her real sessions, taken with the client's written consent. Relevance two, emotion two, authenticity two, space two, clarity one. Nine out of ten, with a release on file. She applies her warm preset and places the text on the right. Her inquiries mention the photo.

4:50 Worked example 2: a travel agency

Worked example two, a business scenario with illustrative details. A travel agency in Islamabad posts destination content and uses images from all over the internet. One day, they receive a copyright claim from a photographer for a mountain photo used in a paid ad. They pay a settlement and fix the process. From now on, every image gets an entry in an image license log: file name, source, license type, date, attribution text, commercial use allowed, and releases. They build a library of their own trip photos, taken by staff and guides with permission from anyone identifiable. For destinations they haven't visited, they license stock with commercial rights. And any AI-generated scenic image is clearly labeled and never used to represent a real hotel or view.

5:45 Watch me do it: scoring three images

Watch me do it. I'm choosing an image for a bakery's weekend post. I have three candidates. Candidate one, a stock photo of croissants from a free site. Candidate two, a phone photo of the bakery's actual croissants on the counter. Candidate three, an AI-generated image of an idealized pastry. I score each. The stock photo: relevance one, authenticity zero. The AI image: relevance zero, because it isn't their product, which would mislead customers. The phone photo: relevance two, emotion two, authenticity two, space one after cropping. I check rights: it's their own photo, no people. I apply the bakery's preset, crop to the left third, and log it. Decision made in two minutes.

6:35 Consistent treatment

Let's talk about treating images consistently, because that's how a series looks like one brand. Set crop and framing rules, like always close-up, subject on the left third. Use the same color adjustments, saved as a preset. Use background removal for product cut-outs on brand colors. Pick a consistent shape or mask, like rounded corners. And use subtle overlays, like a brand-colored gradient behind text. Be honest in editing. Retouching blemishes or cleaning a background is normal. But editing a before and after, or exaggerating a product's effect, can mislead consumers and break advertising rules in many markets.

7:18 Common mistakes

Common mistakes. Using images because they're pretty, not relevant. Generic stock clichés. Assuming anything online is free to use. Ignoring attribution requirements in Creative Commons licenses. Using identifiable people without releases. Leaving competitor logos or trademarks visible in the background. Upscaling tiny images into blurry messes. Using AI images to represent real products or places. And not keeping records. The license log takes thirty seconds per image and can save you a lot of money and embarrassment later.

7:52 Recap and try this now

Let's recap. Choose images with five questions: relevance, clarity, emotion, authenticity and space. Original photos usually beat stock. Know your sources and their licenses, get model and property releases where needed, and keep an image license log. Treat AI images carefully: read the terms, understand their limited protection, keep Content Credentials, and never present them as real. Apply a consistent, honest treatment across your series. Try this now. Create an image license log for the images in your last five posts. Score each with the image scorecard. Replace any image you can't document, and write one consistent treatment rule for your next series.