San Diego, CA tech startup, Truepic, has published an article on its blog that lists several ways to spot AI generated images.
Personally, I tend to look for deformed fingers and mismatched backgrounds. Also, the eyes always seem to be off.
The article states that the most common indicators an image is AI-generated are:
- Unnatural or glossy textures
- Distorted or asymmetrical features
- Background inconsistencies
- Lighting and shadow discrepancies
- Distorted or misspelled words
Honeymoon is over, babe
It appears that the initial wave of infatuation with AI generated media has started to cool down. Especially since the rush to create has resulted in a flood of less-than-mediocre output being vomited onto the interwebs.
Don’t get me wrong. I believe AI has plenty of valuable utility in the right context. I say this from experience.
But something handcrafted continues to retain its superiority – at least desirability – over machine generated goods.
Thankfully, I believe this has led to the increase in value of human-crafted creative works.
Authenticity
The problem is, it’s unlikely that the everyday person out there has the knowledge or resources to peel back layers of deception in order to uncover the real deal.
Like finding an authentic Rolex among a sea of cheap knock-offs.
Truepic thinks this is a problem as well. And they’re building a company based on this belief.
Here’s the company’s tagline and summary from the website:
Visual verification for a digital world.
Understand the authenticity and history of photos, videos, and audio with secure tools from Truepic. Whether you need C2PA compliance or tamper-proof inspections, Truepic enables transparency and trust in all digital media.
New AI Vertical
Truepic isn’t the only player in this space.
FoxNews recently launched their own Verify tool with a keener focus on gatekeeping authenticity within the news information and media sector.
Of course, this raises the question as to who gets to retain authoritative credentials in being the final arbiters of truth. Among other things.
As technology continues to thrive in the metaverse, life in The Matrix is becoming a more plausible scenario.
Discerning what is real or digital is going to become more difficult.
Especially as AI models get better at learning to be more human.