Could AI Headshot Generators Overtake Professional Photographers?

It’s the end of the year 2022. ChatGPT has taken over. The days of writers, programmers, graphic designers, and other professionals are numbered. Get off your computer and go to trade school…at least until AI-powered robots take those jobs over too.

We’re all aware of the hysteria that followed the release of ChatGPT’s GPT 3.5 model. Even now, you can go on Google and start typing “Will AI replace” and get a list of the next apparent targets. As of July 2024, the targets that users suspect are, in order:

1. Programmers
2. Artists
3. Accountants
4. Humans
5. Data scientists
6. Radiologists
7. Finance jobs
8. Mechanical engineers

The results are always changing, but they don’t always follow the real technological transformations taking place. After all, each of these jobs still exist and, at most, have integrated AI into their day-to-day operations.

One of the newest real trends that’s already offering shocking transformations comes from AI image generators. They’ve been around for a while now, but AI image generator algorithms can now offer:

• Thorough machine learning on the intricacies of human faces
• Replication of facial features
• Replication of precise features into different poses and outfits

The new target of the AI revolution is the headshot photographer. Some generative AI startups like Portrait Pal (https://portraitpal.ai/) are making realistic headshots in custom outfits and backgrounds affordable for everyone.

An Alternative to Pricey Photoshoots?

First impressions, including digital first impressions, can make or break one’s professional future. A good corporate headshot has been found to increase trust, represent values, and reflect well on your professionalism. To gain these benefits, people traditionally rely on pricey photoshoots. Even small sessions can be expensive, often prohibitively, for some people.

Given the power that good headshots give you, they aren’t something you want to take an easy alternative for. The money you spend on a great professional headshot can repay its cost many times over.

Most AI headshot generators don’t cost very much. A quick “photoshoot” of between 20 and 50 photos normally costs less than $40.

AI Headshot Generator Quality

There’s a lot of variation in the image quality that AI headshot providers offer you. Not only that, but many AI headshot generators will provide different levels of image quality in one order of headshots.

It’s still hard to get at least one AI-generated headshot that’s usable in a professional setting, and you normally need to make bulk orders to be able to get that. But even this represents the breakneck speed of AI image generation development.

Early Image Generators

When AI was in its beginning phases, it couldn’t create realistic headshots. When you tried, you were likely to see strange deformities and clear glitches in lighting, skin texture, and more. Facial expressions are complex to capture. So early image generators weren’t really capable of producing usable headshots.

CNN

Updates in convolutional neural networks (CNNs) presented one of the most significant turning points in AI image generation. The earliest CNN iterations of the late 90s pale in comparison to even the DALL·E2 or ChatGPT. Newer CNN models provide deeper extraction of more complex surfaces. The recent innovations have unlocked more accurate learning and replication through:

• New optimisers for training techniques
• New data augmentation technology
• Separable convulsions
• Regularisation techniques
• Efficient use of activation functions

GAN

Generative adversarial networks (GAN) innovations started kicking off in the 2010s. This birthed the era of deep fakes, which is where we are today.

While deep fakes are the most commonly reflected on use of GAN in the general culture, there are many subtleties that people don’t consider. While there are of course unethical use cases, GAN may well be the key to democratising art and self-presentation.

In short, GAN, as the “adversarial” implies, sets two neural networks against each other. The supervised learning approach creates more ideal outcomes in a matter of hours. In GAN, the neural networks shoot back at each other in a constant feedback/learning loop.

For example, in AI headshot generation:

1. Network A takes on the role of “generator” and creates a new image based on a subject
2. Network B takes on the role of “discriminator” and compares the original image with the new one
3. Networks A and B take turns generating and discriminating
4. Over time, this generation and discrimination process leads to an increasingly realistic final result

Stable Diffusion

Most recently, stable diffusion made all this technology, and far more, accessible to everyone. With a high-end, but still consumer-grade graphics card, you can start generating ultra-high quality images with AI.

Through this process, companies with the capital to invest in the best hardware can act as intermediaries between consumers and the highest-end and most efficient AI image generators.

Professional Headshots for the Masses?

Innovations in image generators over the years have created the perfect formula for mass access to professional headshots. High-quality office attire and photoshoots can run up costs into the high hundreds or even thousands if you’re using a traditional studio session with a professional photographer. But AI headshot generators are making this process low-cost, or even entirely free!

Using stable diffusion models, companies like PixelPose offer free samples or entire photoshoots at low cost for AI-generated headshots. The process is also fast and convenient, not requiring any work attire or real-world professional headshots of your own. Most importantly, this access comes with high quality and life-like representations of your appearance.

With a few photos or selfies, AI headshot generators can use many features, including GAN, to study your facial features on a deep level. They’re highly self-critical and thus self-correcting, all for generally around $30 to $40 for the apps’ lower tier options. Crucially, they also enable you to realistically transport your face into:

• The professional attire of your choosing
• A professional background of your choosing
• A confident but natural pose

So, will AI headshot generators come for professional photographers’ jobs? In the end it’s hard to say, but the answer is still most likely “no”. Professional photography studios still offer an exceptional level of quality, dedication, and service, which AI tools cannot replicate.

However, for anyone who can’t afford the time or financial commitment of getting an individual or team headshot:

• AI headshot generators have reached a turning point in terms of quality
• Users can gain a realistic photo with a classy outfit and background
• AI headshots are inexpensive and are generated within 3 hours