Decentralized AI gains ground as more protocols collaborate

Lumerin, Morpheus and Exabits are working together to build a blockchain-based AI agent economy, but talent and time constraints still pose challenges.

More artificial intelligence ventures are joining forces to run their services on blockchains, seeking decentralized technology to address AI’s development challenges.

Following the multibillion-dollar merger between Fetch.ai, SingularityNET and Ocean Protocol to counter tech giants’ control over generative AI, now more protocols are seeking similar strategies.

The Open-source protocol Lumerin, for instance, is teaming up with Morpheus and Exabits to promote an “AI agent economy” powered by decentralized computer resources.

While Morpheus connects users with AI services and computing resources, Lumerin will manage and direct the flow of data across the Morpheus network and is behind the core node software. Exabits, a base-layer protocol for decentralized AI computing, backs the computer hardware required for these complex computations.

This decentralized infrastructure will allow what the companies envision as the future of AI, in which agents will be able to perform tasks between Web2 and Web3 ecosystems on behalf of users, making financial services such as exchanging, staking and swapping tokens “as easy as talking to Siri.”

“We are moving into a new paradigm of autonomous economies,” Ryan Condron, Lumerin project leader, told Cointelegraph.

According to forecasts, the newly developing blockchain AI market is projected to grow to $703 million by 2025, with a compound annual growth rate of 25.3%. Issues present in AI development support this expansion. According to researchers at the Massachusetts Institute of Technology, these issues include limited access to data, inflexible models and a lack of transparency and accountability with data and algorithms hidden away.

“Centralized AI models are more prone to inherent biases and increase the risk of censorship and monopoly,” Lumerin said in a statement.

Decentralization of AI could mean users’ context isn’t stored by centralized AI systems, such as ChatGPT or Gemini, and peer-to-peer interactions are more private, but it’s not without its challenges.

According to Condron, startups in this field continue to struggle with time and talent. “Open-source software development for a decentralized network is very, very different than product development inside a company. The challenges of coordination and cohesiveness on the engineering level are hard to navigate in the early stages of these projects,” he noted.

Exabits’ chief marketing officer, Doug Keeney, believes independent AI is essential to building a world that benefits humanity. “Owning our intelligence requires a decentralized approach to AI,” he said.