Inside a Radical New Project to Democratize AI: Autonomous Knowledge and the Future of Knowing

Inside a Radical New Project to Democratize AI: Autonomous Knowledge and the Future of Knowing

In recent years, the development of large language models (LLMs) has been at the forefront of artificial intelligence (AI) research. These models, such as OpenAI's GPT-3 and Google's LaMDA, have shown impressive capabilities in understanding and generating human-like text. However, they have also raised concerns about transparency and access. That's where a radical new project called BigScience comes in.

BigScience is a collaboration of over 1,000 volunteer researchers, coordinated by AI startup Hugging Face and funded by the French government. Their goal is to develop a large language model called BLOOM (BigScience Large Open-science Open-access Multilingual Language Model) that is designed to be as transparent as possible. Unlike other LLMs, BLOOM's creators are sharing details about its training data, development challenges, and performance evaluation.

One of the key selling points of BLOOM is its ease of access. Now that it's live, anyone can download and tinker with it for free on Hugging Face's website. This opens up opportunities for AI developers to use BLOOM as a foundation for their own applications. With 176 billion parameters, BLOOM is even larger than OpenAI's GPT-3, yet it offers similar levels of accuracy and toxicity. This combination of size, accessibility, and transparency makes BLOOM a groundbreaking development in the AI field.

But what sets BLOOM apart from other models is not just its technical specifications. Hugging Face is also taking steps to ensure responsible AI use. They have drafted an ethical charter that guided BLOOM's development and are launching a Responsible AI License. This license acts as a deterrent from using BLOOM in high-risk sectors or for harmful purposes. By prioritizing ethics and transparency, Hugging Face is setting a new standard for the AI community.

In parallel to the democratization of AI, there are discussions about the future of knowing itself. J. Adam Carter, in his work "Autonomous Knowledge: Radical Enhancement, Autonomy, and the Future of Knowing," explores the concept of knowledge beyond justified, true, non-Gettiered belief. Carter argues that knowledge also requires autonomy, which cannot be reduced to other components.

Carter's argument revolves around the idea that relying on external devices for information acquisition hinders true knowledge. Even if we have justified, true beliefs acquired through these devices, they lack autonomy because they were obtained heteronomously. Carter believes that true knowledge requires a causal history that is free of compulsion.

However, critics may challenge Carter's autonomy condition, suggesting that autonomy should be a condition on justification rather than knowledge itself. Others may argue that subjects can possess know-how even if the skill is caused by a mechanism they don't own, as long as it is integrated with their abilities.

Despite these objections, Carter maintains that autonomy is crucial for both propositional knowledge and know-how. He suggests that sheddability, the ability to easily shed a belief, is a key factor in determining autonomy. Additionally, Carter argues that innate faculties, whether natural or implanted, can be relied upon in the same way.

The intersection of democratizing AI and the future of knowing raises important questions about the responsible and ethical use of emerging technologies. Here are three actionable pieces of advice to consider:

1. Prioritize Transparency: When developing AI models, strive for transparency and openness. Sharing details about training data, development challenges, and evaluation methods can contribute to the overall trustworthiness and accessibility of AI systems.

2. Embrace Responsible AI Practices: Establish ethical guidelines and codes of conduct to guide the development and use of AI technologies. Consider implementing licenses or agreements that discourage their use in high-risk sectors or for harmful purposes.

3. Foster Autonomy in Knowledge Acquisition: When exploring the future of knowing, recognize the importance of autonomy. Strive to understand how external devices, such as neural interfaces, can be integrated in a way that maintains autonomy in knowledge acquisition.

In conclusion, the democratization of AI and the concept of autonomous knowledge are two intertwined developments shaping the future of technology and human understanding. By prioritizing transparency, responsible AI practices, and autonomy in knowledge acquisition, we can navigate these advancements in a way that benefits society as a whole.