2019 was characterized by cautious innovation in language models, alongside significant breakthroughs in image manipulation, audio synthesis, and AI ethics. This year highlighted the delicate balance between pushing the boundaries of AI capabilities and addressing ethical concerns surrounding powerful AI technologies.
November
- OpenAI fully releases GPT-2
- Decision made after research indicated no significant malicious use
- Marks a milestone in responsible AI deployment
June
- Nvidia releases GauGAN
- AI-powered tool that turns simple sketches into photorealistic landscapes
August
- OpenAI releases a partial version of GPT-2
- Released version has 774 million parameters
- Full version initially withheld due to concerns about potential misuse
January
- IBM unveils Project Debater
- First AI system capable of debating humans on complex topics
- Demonstrates ability to formulate coherent arguments and rebuttals