AI Meetup Yesterday

I did another AI meetup yesterday and thought I'd send out some topics and technologies I found interesting.

Apparently you can check the provenance of these types of files:

• Image: AVIF, DNG, HEIC, HEIF, JPEG, PNG, SVG, TIFF, and WebP.
• Video and audio: AVI, M4A, MOV, MP3, MP4, and WAV.
• Document: PDF.

So I made this image through o3 model ChatGPT interface:'

It recognized it as originating from OpenAI. So it worked on that example.

I gave Canva a try and it didn't seem to work there with a saved design.

However, for Canva it didn't work. It didn't recognize any records from where it possibly originated from.

I'd imagine it might work on some software that originates from Adobe like Photoshop. However, I don't really have any Adobe products or services. I'm basing that from the logo board they have on their site.

What this means for the files you handle today

1. Most ordinary web images still carry only EXIF/IPTC, and many social platforms strip even that.
2. C2PA-signed files are climbing quickly because Adobe now ships Content Credentials on by default in the web app beta and is pushing partners to respect the manifest.
3. Invisible watermarks complement, not replace, metadata—they prove origin but don’t tell the story of edits or licensing.
4. Provenance in software artifacts is becoming table-stakes for regulated industries (e.g., U.S. government’s secure-by-design guidance) thanks to SPDX + SLSA.

Bottom line: if you’re publishing or ingesting digital assets in 2025, you’re likely to encounter at least one of these provenance layers. For robust authenticity you’ll want to (a) preserve existing metadata, (b) verify C2PA manifests where present, and (c) emit your own signed provenance when generating new content.

MCP helps you build agents and complex workflows on top of LLMs. LLMs frequently need to integrate with data and tools, and MCP provides:

• A growing list of pre-built integrations that your LLM can directly plug into
• The flexibility to switch between LLM providers and vendors
• Best practices for securing your data within your infrastructure

I made a group chat to discuss this topic:

According to a post by Gareth Manning, graphic designers have one of the lower automation risks: https://garethmanning.substack.com/p/high-tech-low-literacy-how-the-ai

I'm not calling that a credible source or a rigorous analysis, but I do think that it's interesting to see that people somehow feel that certain professions are safer or riskier to being replaced. Each person has their own rationale as to why some jobs are riskier or some jobs are safer.

For example, that post's author Gareth, he's got a background in education it seems.

Gareth also puts secondary school teachers as a profession unlikely to be replaced.

However, has Gareth ever played around with Khanmigo? Has he seen some of the robotics companies that people have created in recent years?

My thinking is that, what happens to artists will be what happens to most people as automation technologies become more prevalent. So it's worth a place to discuss it.