Artificial intelligence (AI) has the potential to usher in a new era of tools and technologies that will benefit people and their communities, creating new economic opportunities and accelerating creativity and innovation.
It also has the potential to do the opposite, building walls around closed-loop systems that strengthen existing corporate interests, eventually shutting out Web3 forever.
Chris Donovan, chief operating officer of Near Foundation, the governing body of the blockchain abstraction protocol, Near Protocol, told Cointelegraph what is troubling him and his fellow colleagues.
“Our great concern right now when talking about the open web is that if we don’t have a credible, irrevocably open environment for AI development, then the open web may never come to fruition,” said Donovan.
Illia Polosukhin, co-founder and CEO of Near Foundation, explained to Cointelegraph why AI poses a risk to open systems.
“Most AI development is happening within really large corporations which have a very clear incentive where they need to continue generating more revenue for the company,” said Polosukhin.
Polosukhin calls this phenomenon “corporate-owned AI,” and the threat he describes isn’t immediately obvious. But Polosukhin, who himself worked at Google for over three years, goes on to explain the matter further.
Recent: Iceland’s crypto industry has promising future, but faces energy challenges
“This [AI] has very antagonistic goals to the individuals that are using it because Google or Facebook can’t find another billion users somewhere,” he says. “They just don’t exist on the internet. So they need to grow their revenue per user, and to do that, they need to get users to spend more time on their product.”
In this scenario, large corporations utilize AIs to corral users into corporate silos.
“It’s a very dystopian future if you imagine that,” says Polosukhin.
Closed doors and manipulation
Polosukhin looks to lessons from the past to inform his prediction of the future.
“If you assume the current wave of corporate-owned AI is continuous, at some point, you are just sitting on an operating system or on a social network, and you literally don’t have any other way out. You’re not going to even have a browser,” says Polosukhin.
Polosukhin points to Microsoft, a leader in the AI race, which has, at various times, pursued anti-competitive practices. In 2001, the United States brought an anti-trust lawsuit against Microsoft for attempting to monopolize the web browser market.
The case was eventually settled when Microsoft agreed to amend business practices that were deemed unfair.
Microsoft’s approach to stifling competition would appear to have changed little in the intervening years.
On June 25, the European Union charged Microsoft with unethical practices related to its bundling of Microsoft Teams with Office. According to EU regulators, this software bundling gave Microsoft an undue advantage over competitors such as Slack and Zoom.
But while these examples are damning in and of themselves, Polosukhin believes that AI could create a perfect storm of anti-competition because of its potential to actively manipulate its users.
“You will not be able to escape that kind of system, and you will think that’s actually what you want,” said Polosukhin.
Social ills and social media
While the monopolization of software and media platforms is one risk factor concerning Near’s founders, Donovan predicts that AI could accelerate social media’s worst tendencies.
“These platforms don’t incentivize good behavior or high-quality information. They incentivize attention and eyeballs on the screens, which I think is a very negative paradigm,” says Donovan.
“I think if AI is deployed in that paradigm, it will just amplify and accentuate all of those issues — and those issues are actually really fundamental. They’re things like political polarization and erosion of trust in institutions and misinformation and disinformation, and all of these types of things that are just damaging our society on a global level.”
Polosukhin further believes that large tech platforms will want to provide more and more information directly rather than diverting to third-party sources, eventually diluting the plurality and independence of media.
“You will not leave Google. You’ll not go to smaller pages. You will not give them any revenue anymore. All the publications [...] will have less and less traffic from Google,” said Polosukhin.
The potential for AI tools
Having discussed negative scenarios at length, Cointelegraph asked Polosukhin to discuss some potential advantages AI can bring.
Polosukhin said that AI has the potential to make individuals more successful and transform even non-tech-savvy users into builders.
“One of the main things we are working on is the AI developer — this ability for individuals to build applications from scratch. Even if you’re non-technical, you can come in and build or customize existing applications,” says Polosukhin.
“Maybe somebody has built a tool for you, but it probably has a bunch of things that you would want to change. Imagine that while you’re using it, you press tab, and you can add a new feature,” outlines Polosukhin. “That’s how you become more productive, how you get more empowered and how you build the environment around you that actually suits you best.”
Another area Polosukhin highlights is AI assistants that can complete tasks such as booking flights or other activities.
Unsurprisingly, for a firm with such strong opinions on the future of AI, Near is pursuing its own vision of artificial intelligence, which it calls “user-owned AI.”
At the launch on June 20, Near announced six partners for the project spanning data, models, compute and inference.
Recent: Celebrity memecoins highlight crypto’s influencer problem
According to Near, these solutions working in concert will help create the foundation of an AI ecosystem that is “secure, decentralized, and tailored to empower end-users as well as researchers contributing to open source.”
According to Polosukhin, the vision for Near is to become the “underlying layer” while “the protocol itself is the identity, the payments” that facilitates the grander design.
“You’re not going to run AI inference onchain,” clarifies Polosukhin. “You run it offchain on specialized compute [...] but we are building a few different components that tie all of this together to enable this future.”