Jipitauri Georgian Language Chatbot (GPT 3.5) Use Cases and Challenges

Transcript

1. Pulsar AI Georgia Supernova
2. 01 Pulsar AI was established in 2017 Products, Pulsar AI started with the first Georgian language bots with powerful components of AI
3. AI, Big data analyses, Process automatization 02 Real Time data analysis Transaction Validation/Control Image Processing Payment Fraud Prevention Many more
4. 03 Trust in AI Responses The Bias Problem Ethical Behaviour The amount of Data Skills Required Challenges in testing AI products
5. 04 • Customer support • Personal assistant • Language learning and translation • Content generation • Code writing and debugging ChatGPT, Generative AI and GPT-3 Apps and use cases
6. 05 Chat GPT 2873 individuals evaluated the 7675 texts produced by human and GPT-3 survey respondents Department of Political Science, Brigham Young University
7. 06 Jipitauri based on GPT 3.5 Georgian ChatGPT :) Model was enriched with Georgia-specific, historical and current facts
8. 07 Jipitauri Highlight the topics and sentiment Developers, We got a very pure result, identified top 5 topics, overall sentiment was negative due to the incomplete data
9. 08 Jipitauri Highlight the topics and sentiment https://chat.supernova.ge/
10. 09 GPT-4 is the next iteration of the language model series by OpenAI Again, Challenges Ahead • Bias • The Challenge of Computational Power • The Challenge of Data Privacy Developers, researchers, and society have to ensure that they use AI language generation technology in an ethical and responsible manner
11. en.supernova.ge Keti@pulsarai.ge Thank you