Experts Say Decentralized AI Offers New Options for Data Control

As part of the Blockchain Futurist Conference in Toronto, panelists convened to discuss the impact of artificial intelligence (AI) and blockchain on people’s lives.

During the “Future of AI and Blockchain” panel, discussions surrounded the nature of centralized AI, how blockchain can decentralize AI processes and the future of AI. As AI becomes increasingly controlled by a select number of corporations, panelists voiced skepticism over AI’s current trajectory.

Panelists included Toufi Salibi, serial entrepreneur and CEO of Toda.Network; Chantel Costa, director of development at the DAIA Foundation; and Dr. Ben Goertzel, CEO of SingularityNET.

“AI systems are interwoven with data about all of us and with the actions that we take in all of our lives,” Goertzel said. “What this means is that who is controlling and guiding these AI systems becomes of critical importance.”

As a small number of large corporations and governments control more data, Goertzel suggested that decentralized technologies, such as blockchain, offer an alternative system to control data and AI systems.

Goertzel added that the majority of AI systems are run by Microsoft (NASDAQ:MSFT) Azure, Amazon (NASDAQ:AMZN) AWS, Google (NASDAQ:GOOG) Cloud and Tencent (OTC Pink:TCEHY) Cloud.

The control of this data does not necessarily have to be confined to these few large players, he added.

The discussion then led to distinguishing between two forms of AI: narrow AI and general AI. Narrow AI refers to AI systems that are trained on a specific, small task. General, or strong, AI acts closer to a human, with a deeper understanding of the world. General AI has not yet been achieved.

Goertzel added that the increasingly concentrated control of AI by large corporations will pose a bigger problem as they shape the future of AI.

“As today’s narrow AI systems verge into tomorrow’s artificial general intelligence systems, it’s going to become more of an issue,” he said.

Salibi, who spoke in the panel about how blockchain is inherently designed to reduce friction, addressed the issue of centralized AI while also advocating bringing data under its owner’s control. “Everything that belongs to you, stays with you,” he said.

“The intent of the technology is to reduce the friction tremendously from where it is today,” Salibi said. Despite blockchain’s 10 year history, the penetration rate of blockchain only stands at 2.2 percent, he added.

In terms of AI’s influence on democracy, Costa suggested that, while there isn’t democracy in AI systems, there is promise.

As an elite group of companies acquire and control vast amounts of data and AI talent, it may necessitate the need to call for data regulation in addition to new solutions.

“This is where blockchain could come into the picture,” Costa said.

Naturally, the topic surrounding the efficiency of decentralized AI systems also came to the surface. Panelists agreed on the need to reduce friction in blockchain technology to better leverage its utility and potential.

Goertzel acknowledged that, currently, centralized AI systems operate more efficiently than decentralized systems. Countering these problems, systems like Toda.Network are working to create more efficient centralized systems, he added. He said these systems also have features within them, including diversity and creativity.

“Given the greater ability to attain diverse inputs, serve diverse needs and then foster creativity by the cross-pollination of interaction of diverse AIs, … this can make a decentralized network much smarter, … because it is not maximizing shareholder value for a few corporations,” said Goertzel.

Don’t forget to follow us @INN_Technology for real-time news updates!

With files from Georgia Williams.

Securities Disclosure: I, Dorothy Neufeld, hold no direct investment interest in any company mentioned in this article.