[Excerpt from manuscript chapter 3]
The research landscape of artificial intelligence and artificial general intelligence (AGI) is characterised by a high degree of dynamism and diversity. A number of players and institutions can be identified that are making significant contributions to the further development of this discourse. Various organisations and academic institutions through to private research laboratories are continuously pushing the boundaries of what is technologically possible and laying the foundations for the development of systems with general intelligence.
The International Telecommunication Union (ITU), as the UN specialised agency for information and communication technologies, is one of the most important players in the research landscape. The ITU leads the ‘AI for Good Initiative’, a global platform that promotes the contribution of AI to achieving the United Nations’ Sustainable Development Goals.
At the annual ‘AI for Good Global Summits’, practical applications of AI are identified to accelerate progress in areas such as education, healthcare and socio-economic equality. The initiative aims to create the necessary partnerships to bring promising ‘AI for Good’ to fruition on a global scale.
ITU has also entered into a partnership with the AI Commons platform, a network that promotes solutions, datasets, models and resources for scaling AI projects to improve social and economic interests. The members of this network include experts such as Yoshua Bengio, Primavera De Filippi and Yann LeCun.
Under the direction of Yoshua Bengio, Mila (Quebec AI Institute) is one of the world’s leading research institutes in the field of machine learning and AI. In addition, Yann LeCun’s laboratory at Meta (formerly Facebook) plays a central role in AI research. Yann LeCun, one of the pioneers of deep learning, leads Meta AI’s efforts to develop advanced AI models and technologies. Research focuses on areas such as self-supervised learning, computer vision and the development of efficient neural networks. Meta AI has made significant progress in the development of models such as the open source language model OPT (Open Pre-trained Transformer) and actively contributes to the wider AI community.
OpenAI made headlines in 2022 with its GPT-3 (Generative Pre-trained Transformer) AI model for natural language processing. With over 175 billion parameters, it was the most powerful system of its kind to date and the starting point for other multimodal AI models such as DALL-E for text-to-image generation.