A brief overview of the development of Generative AI

Weijia Liu, Miaomiao Li

Article ID: 8583
Vol 7, Issue 8, 2024

(Abstract)

Abstract


Since the 1950s, with the development of computers and information technology, especially the breakthroughs in deep learning
technology, Generative AI has gradually evolved from a single language generation to a multimodal and embodied direction. And has
shown a strong vitality through pretrain and prompt, with data-driven and powerful generative capability, deep learning and highly
customised, intelligent simulation and real-time efficiency. It has the characteristics of data-driven and powerful generative ability, deep
learning and highly customisable, intelligent simulation and real-time efficiency. A clear overview of the development of Generative AI will
help promote agile governance and the development of technology for good while empowering various industries.

Keywords


Generative AI; Pretraining; Multimodal

Full Text:

PDF


References


1. [1] Xiaoli Wang,Chi Yan. Risk issues and regulatory progression of Generative AI macromodels: an example of GPT-4[J/OL]. Journal

2. of Beijing University of Aeronautics and Astronautics (Social Science Edition),1-11[2024-06-20].

3. [2] Wen Zhao. Research on Generative AI Generated Content Governance - A Perspective on ChatGPT[J]. Communication and Information Technology,2024,(02):95-99.

4. [3] RYAN T. Chips and Science Act [R/OL]. (2022-08-09)[2023-04-05].

5. [4] David Rozado, “The Political Biases of Chat GPT,” Social Sciences, Vol. 12, No. 3, 2023, pp. 1-8.

6. [5] HU Bing,JONG Siyue. Implications of the Dutch preemptive techno-ethical thinking for AIGC governance[J]. Young Journalist,2024(02):48-54.

7. [6] Zhang Haopeng,Xia Baohua. What makes artificial moral intelligences feasible - based on an examination of value-sensitive

8. design[J]. Natural dialectics research,2021,37(04):37-42.




DOI: https://doi.org/10.18686/ijmss.v7i8.8583

Refbacks

  • There are currently no refbacks.




Creative Commons License

This site is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.