AIs talk to each other

AI Blog

Issue #29

Synthetic Humans or how AIs talk to each other

We don’t know if you were expecting this moment, but here it is. Fantasy (AI Agency) has created Synthetic Humans.

Synthetic Humans are AI chatbots designed to mimic human behavior and conversation, trained on over 100 attributes of the human psyche. Each has an extensive backstory, personality, and characteristics.

Their primary purpose is for use in market research, participating in focus groups either alone or with humans. They can also function as team members and assistants.

Behind these chatbots are LLMs (Large Language Models) like GPT-4, Claude, etc., and according to Fantasy, they can “see” images.

LLMs are generally very good at mimicking human “speech,” but such anthropomorphism requires great caution. We only need to recall the case of Blake Lemoine (discussed in Issue 7, read more here.)

Additionally, Synthetic Humans remind us of the business version of the Character Group Chat released by For those who want to study the subject further, we recommend reading John J. Horton’s research on Homo Silicus, as he calls simulated humans powered by AI.

Adobe levels-up its AI products 

Last week, Adobe MAX 2023 took place, where Adobe announced various updates related to AI & creativity, including new AI models and AI-enhanced features of existing ones like Photoshop.

Firefly 2, the improved version of Firefly, can now generate higher-quality images and illustrations, with improvements in human features (e.g., eyes, hair, skin color, etc.), allowing for better output control.

Over 100 new AI features have been added to all Creative Cloud applications. Noteworthy mentions include Illustrator, Photoshop, and Lightroom. Additionally, Adobe Express has been enhanced with AI capabilities like Generative Fill, Text to Template, and Translate, aiming to provide a more seamless user experience.

Adobe remains the most business-friendly option in the AI sector for creativity and a pioneer in addressing copyright & ethics issues. It offers Content Credentials, a “digital tag” (metadata) attached to AI-generated assets, promoting greater transparency around AI-generated content.


Bonus, some interesting statistics:

  • 3 billion images have been produced with Firefly
  • 90% of users are net-new to Adobe products.



AI Insider 📰

  1. Disney is once again under criticism for its use of AI. The issue arises with the background of the poster for season 2 of the series Loki, which appears to be an AI-generated image from Shutterstock. According to Shutterstock, AI-generated images cannot be licensed unless created with its own AI tool.
  2. ElevenLabs κhas released a new tool for voice dubbing. It can translate into 20+ languages (including Greek), maintaining the original voice and making it sound as natural as possible. This follows similar tools tested by  YouTube and Spotify with their creators.
  3. Η Replit has announced its new suite of AI-powered coding tools, which will be open access. Their model, code-v1-3b, can be found on HuggingFace.
  4. CapCut will roll out AI-generated ad scripts and product presenters.
  5. A bit more gossipy, but The Information published an article yesterday about OpenAI’s project ‘Arrakis,’ which had to be discontinued as it did not meet their expectations.

Learning Bytes 🧐

  1. US lawmakers want to pass the No Fakes Act in response to AI-generated songs with artists’ voices flooding TikTok. The official announcement states that the No Fakes Act “would prevent a person from producing or distributing an unauthorized AI-generated replica of an individual to perform in an audiovisual or sound recording without the consent of the individual being replicated.” The phrase “without the consent” is crucial here, as celebrities already rent their voice & image to companies for AI use.

  2. Google announced early access to Vertex AI Search for companies in the healthcare sector to facilitate the search for necessary data.

  3. Anthropic published research attempting to explain LLMs in a more understandable/simple way (compared to the analogy with the nervous system).
  4. Geoffrey Hinton, known as the godfather of AI, gave an interview on 60 Minutes, where he talked about how in 5 years, AI might reason better than humans. You can watch it here. 
  5. Researchers have developed AI that can predict future mutations in viruses.

Cool Finds 🤯

  1. For anyone wanting to express themselves through creative writing, the prompt of the month in the Six-Word Sci-Fi: Stories Written by You is “AI on trial.”
  2. Additionally, the Sydney Opera House presents Music of the Sails. This is a livestream lasting throughout October, where we hear the daily data received by the Opera House processed by AI to recompose into a unique musical soundscape. The music is accompanied by AI-generated images. The project is part of the celebration of the Opera House’s 50th anniversary.



See you next week! 💚




Social Media Awards



E-volution Awards



Peak Performance Marketing Awards



Effie Awards