Talking with machinesCREATIVITY THANKS TO ARTIFICIAL INTELLIGENCE?
- 3 min reading time
In December 2022, ChatGPT 3 was made available to the general public and the internet community. Within 5 days, the voice AI model, which is able to conduct human-like conversations and provide answers to questions, broke the 1 million user:in barrier. In the course of this, AI image generators such as Midjourney or Dall-E-2 also became more widely known and triggered a real AI hype in the creative industry.
VHUG & AI
Already at the end of last year, we experimented with AI in the course of our Christmas card: With the QR code printed on the blank card, an augmented reality application can be opened in which around 20 headlines on the topic of Christmas are available.
If you select one of the headlines, AI-generated Christmas motifs are played - from the classic Christmas tree with "Bethlehem is calling" to the absurd "Reindeer Shit Bingo".
Innovation or just the next hype?
Tools like ChatGPT or Midjourney promise creativity and results at the push of a button. Texts, images and illustrations, concepts, music - whatever your heart desires can be created in a few seconds. However, the result still varies extremely. In addition to the lack of consistency and transparency of these tools, one's own "prompts" also strongly determine the quality of the results. However, if you consider that this technology is still in its infancy, you understand the hype around ChatGPT and image generators like Midjourney a little better.
It also explains why we have to continue to face this topic with open eyes and ears. We have compiled our first learnings and experiments for you here and presented them in this article.
✅ Increased efficiency and productivity in creative tasks.
✅ More opportunities for experimentation and iteration in the creative process.
✅ Automation of recurring tasks.
✅ New tools and workflows can be developed.
✅ Improved ability to generate multiple ideas.
❌ Bias in AI algorithms.
❌ Risk of diminishing the value of creativity.
❌ Incorrect information being disseminated.
❌ How and with what data was the AI trained?
❌ Copyright infringements? (lawsuit in the USA)
Potential Use Cases for ChatGPT
✅ Online marketing
✅ Concept & Research
Unlike ChatGPT, these tools are not based on text-to-text but on text-to-image. The AIs were trained with an enormous image data set and continue to create new "remix creations" from this and the new compositions. This procedure, the possible infringement of copyright and the monetisation of these tools, is currently the subject of controversial discussion. In addition, the results are strongly dependent on one's own "prompts" and vary greatly in result and quality. It's fun in any case, even if your own creations sometimes remind you of wild fever dreams.
Tools to try:
✅ Dall-E 2
✅ Stable Diffusion
Possible Use Cases for Image Creation AI's
✅ Stock Photos & Photography
✅ Character Design
(All Images in this Blogpost are AI created)
Ready for the next level?
We continue to follow the latest developments in the AI industry and test the latest tools - for and with our customers.
VON HELDEN UND GESTALTEN GmbH