Can you tell if something is AI generated?

 The three images above show similar winter scenes of aspen trees. Looking at them, which one do you think is AI-generated, and which one came from my camera (specifically, my iPhone 17)? Are they all generated, or are they all original photos? What do you think?

There are many YouTube videos claiming to offer foolproof ways to identify AI-generated images. While it is certainly possible to detect AI by examining the underlying code, doing so from a practical, visual standpoint is becoming increasingly difficult—if not nearly impossible. This is especially true when comparing AI to images taken with late-model cameras that use sophisticated internal settings and computational adjustments.

To test your eye, I will reveal that one of these images was taken with my iPhone, one was generated by Google Gemini, and the third was created using Adobe Photoshop.

Paragraph One

I am aware of the controversy that is raging across the academic community in all levels about the use or abuse of AI.  There are persuasive arguments on both sides: the adoption of AI wholesale or banning it completely from the academic setting. However, just as with the three images, it the AI is as sophisticated as Google Gemini, there is no real way to tell whether writing was generated by AI or written by me or someone else without using AI. One reason for this is that I use Google's spell and grammar checker. To the extent that these Google tools employ AI, everything I write in my blogs has been altered by AI at least as to spelling and my propensity to drop out words. 

Paragraph Two

There is currently a significant debate raging across the academic community regarding the use, or potential abuse, of AI. On one side, you have the early adopters who believe we should embrace these tools wholesale to speed up the tedious parts of our work. On the other, there are those who argue for banning AI entirely from academic and historical settings to preserve "authenticity."

What do you think?

The genealogical landscape is shifting beneath our feet, and it’s doing so at a pace that can feel a bit dizzying for those of us who remember the days of cranking microfilm readers in dimly lit archives. For decades, the "gold standard" of research was defined by the physical trail—hand-written notes, photocopies of census records, and the slow, methodical process of manual verification. But today, the conversation has moved from the filing cabinet to the cloud, and specifically, to the role of Artificial Intelligence in our research and writing.

This is largely because the transition to AI isn't always a deliberate "choice" to let a machine write for us; it’s baked into the tools we already trust. For instance, as I mentioned above, I rely heavily on Google’s spelling and grammar checkers. To the extent that these tools employ AI to suggest better phrasing or catch my tendency to drop words when I’m typing too fast, nearly everything I publish has been touched by AI.

The challenge for the modern family historian isn't to fear the technology, but to master it while maintaining our traditional standards for accuracy and sourcing. If you’re ready to integrate these tools into your workflow without losing your "genealogical soul," here is how I recommend approaching it:

• View AI as an Editor, Not an Author: Use tools like Gemini or specialized grammar checkers to clean up your prose and catch those pesky typos, but ensure the underlying facts and citations remain your own.

• Verify the Source Material: Never let an AI summarize a record for you without checking the original image. AI can hallucinate dates or names; your eyes on the original document remain the final authority.

• Be Transparent: If you use AI to significantly restructure a family history narrative, a simple footnote or disclaimer is a great way to maintain the integrity of your work, much like we cite our traditional sources.

We are entering an era where the "how" of our writing is changing, but the "why"—the pursuit of truth in our family lines—must remain the same. The tools are getting smarter, and it’s our job to ensure our research methods keep pace.

Obviously this blog post contains AI generated text. By the way the middle photo came from my iPhone the other two were generated by gemini, the top photo, and Photoshop, the bottom photo. Paragraphe Two is a rewrite of the first paragraph by Gemini as are the rest of my ideas in the last paragraphs. I use Google Gems to draft prompts that incorporate my own style and tone of writing. Google has over 6000 examples of my blog posts to work from. 

In just three short years, AI has become so pervasive as be ubiquitous. There is really to way to avoid it. As far as the academic community is involved in this AI revolution, they will either figure out ways to teach in the AI world or disappear as irrelevant.