Who's Who
in AI

The SAFE AI Foundation Who's Who in AI is a list of inventors who have made significant advances to the field of AI. Without these inventions, there would be no AI, no AI products, no AI business, etc. We do not include AI business leaders or AI investors here. For Top AI voices, please refer to LINKEDIN.
If you think there is an important inventor that should be included, please contact us at contact@safeaifoundation.com with details of your proposal. We reserved the final right to include or exclude anyone after consideration.
Who invented Neural Networks?
1943 - Warren McCulloch & Walter Pitts
Who invented Perceptron?
1957 - Frank Rosenblatt
Who invented Deep Learning?
1965 - Alexey G. Ivakhnenko
Who invented NLP?
Allan Turing, Joseph Weizenbaum, Noam Chomsky
1966 - ELIZA Chatbot by Weizenbaum
Who invented Backward Propagation?
1970 - Seppo Linnainmaa - Published efficient reverse-mode automatic differentiation, the modern foundation of backpropagation
1974 - Paul Werbos 1974's PhD Thesis proposed using backpropagation to train neural networks
Who invented RNNs?
1920s - Wilhelm Lenz and Ernst Ising - First mathematical recurrent architecture (Ising model)
1943 - Warren McCulloch and Walter Pitts - Discussed neural networks containing feedback loops
1972 - Shun-ichi Amari - First published learning RNN
Who invented Hopfield Networks?
1982 - John Hopfield
Who invented Jordon Networks?
1986 - Michael Jordan - Recurrent architecture using output feedback and context units.
Who invented LSTM?
1986 - Sepp Hochreiter and Jürgen Schmidhuber
Who invented CNNs?
1975 - Kunihiko Fukushima
first CNN called Cognitron 1975, 1979
Who invented "AlexNet"?
2012 - Ilya Sutskever, Alex Krizhevsky
Who invented GANs for GenAI?
1991 - Jürgen Schmidhuber - Early generative-adversarial principle (Artificial Curiosity)
2014 - Ian Goodfellow and Yoshua Bengio - Modern formulation of GANs for AI
Who invented Diffusion Model for GenAI?
2015 - Jascha Sohl-Dickstein
Who invented the Transformer Architecture in AI?
1991 - Jürgen Schmidhuber - Earlier unnormalized linear Transformer mathematical principle
2017 - Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan Gomez, Lukasz Kaiser, and Illia Polosukhin -The Google team's original transformer architecture has introduced:
- Scaled dot-product attention.
- Multi-head attention.
- Positional encoding.
- Encoder-decoder architecture.
- Softmax-normalized attention.
The Era of AI
Embracing AI for a better quality of life.
Show you care!
Support us by entering your email. It is free to join as a supporter.
contact@safeaifoundation.com
Email:
© 2026. All rights reserved.
A non-profit 501c(3) organization registered with the State of California, USA




