Written By: Alan Glazier, OD, FAAO, Dipl ABO.
In 2022, ChatGPT, an Artificial Intelligence(AI) enabled chatbot, was released to the public. Each version that came out was smarter than the previous version, and more advanced competing models continue to be developed. Along with these advances came doomsday predictions that AI would achieve a “superintelligence” and ultimately reach “sentience”, a point at which the AI grows beyond human ability to control it. What concerns most people familiar “in the know” is not AI reaching sentience, that is the concern, but who controls it. Many experts believe that sentient AI, in the hands of a centralized authority such as a business or government, could be used to achieve its agendas at the expense of society or for other nefarious purposes. This might be the biggest fear we face from AI. (1,2)
A movement that seeks to take AI out of the hands of centralized organizations and “decentralize” it exists and is gaining steam. Bittensor, a network existing at the intersection of cryptography and AI has a mission to democratize AI and is growing at an exponential rate. Bittensor is a transparent, open-source network that allows collaboration and contributions from anyone developing an AI tool on the platform. It is not a company and it is not owned by anyone. Within the network, developers create AI applications known as “subnets” and compete for technological and functional dominance in their respective verticals. Developers are financially incentivized to create better models than competitors, driving competition and innovation. The tools and documentation used to build and interact with subnets are available to anyone on the network. If one subnet beats out another subnet in terms of output and performance, it financially benefits while the other model loses or dies out. Subnets can compete or collaborate to advance their AI technologies if they choose.
AI experts are watching this platform closely, as many see Bittensor becoming the foundation of what ultimately will dominate the AI space: decentralized AI by the people, for the people, taking the power to control technology out of the hands of corporations and governments. Every day, more entrepreneurs launch subnets. Every subnet deepens and increases Bittensor functionality. Every improvement makes the network smarter, not only creating a better AI but a distributed one that is not controlled by any one entity. I believe it will become the “rails” on which decentralized AI evolves and ultimately surpasses any one centralized platform. Decentralized development, like Bittensor, is the best way to keep us safe from the threats of control we fear from centralized AI developers, freeing developers and incentivizing them to drive AI development forward without constraints that might limit AI abilities. The success of Bittensor or other decentralized networks is vital to ensure governments or for-profit companies acting out of self-interest don’t continue to lead us towards the AI “dark side”, instead of evolving the technology for the benefit of society.
References and further read
- Potential Future AI Risk 5 Page 22, OECD, Assessing Potential Future AI Risks, Benefits and Policy Imperatives, AI Papers No. 272, Nov. 2024
- https://aistatement.com/
- https://bittensor.com/
