Artificial intelligence (AI) and blockchain are emerging digital technologies that have captured the public imagination but also raised serious concerns.
So it’s worth asking: Can artificial intelligence and blockchain merge in a way that benefits humanity?
There are reasons to think so. Back in 2016, Vitalik Buterin wrote Both the cryptoeconomics and AI security communities are “trying to solve essentially the same problem,” namely how to regulate complex and intelligent systems with “unpredictable emergent properties.”
After all, both rely on the control of an inherently “dumb” system “whose properties become inflexible once created.” For example, once a smart contract is implemented, it cannot be changed. The two communities “should listen to each other more,” he concluded.
Over the past year, with the emergence of ChatGPT and other generative AI tools, concerns have grown that artificial intelligence could spiral out of control. In a nightmare scenario, humans could lose control of autonomous weapons systems.
Therefore, the idea that blockchain and smart contracts can somehow act as guardrails to stop AI models from going off the rails has become increasingly popular.
“Everyone working in the crypto space has a unique role to play in making AGI a success.” explain Foresight Institute President Allison Duetmann at the recent SmartCon 2023 conference. This is especially true given that artificial general intelligence (AGI), where machines achieve human-level intelligence, may become a reality. future Better sooner rather than later.
IT decision-makers who participated in a recently released survey also considered this potential convergence of artificial intelligence and blockchain technology entrust Courtesy of Casper Laboratories. Nearly half (48%) of 608 IT leaders surveyed in the U.S., Europe and China agreed that “the integration of artificial intelligence and blockchain technology has the potential to revolutionize our industry to enhance data security, transparency and efficiency” .”
Complementary technologies, growth momentum
The basic idea is that blockchain’s immutable, tamper-proof ledgers, together with smart contracts, can provide guardrails for the implementation of artificial intelligence and ensure responsible artificial intelligence. Some believe that blockchain could even serve as a kind of “kill switch” for runaway AI models.
In a survey commissioned by Casper Labs and conducted by Zogby Analytics, 71% of IT leaders said they “view blockchain and artificial intelligence as complementary technologies.” Additionally, when asked how their organization is currently using blockchain, “Overall, working efficiently with artificial intelligence was the most popular response (51%).”
Elsewhere, November 1, US President Joe Biden release Executive order establishing new artificial intelligence safety standards. The order is designed to protect the public from a variety of risks, including dangerous artificial intelligence-engineered biological materials, artificial intelligence-driven fraud and deception.
Mrinal Manohar, CEO and co-founder of Casper Labs, said in an interview with Cointelegraph that the order “creates huge momentum.” Casper Labs has an enterprise-focused layer 1 blockchain. Today, artificial intelligence governance has become the focus of more enterprise IT personnel.
Does he see more companies launching actual AI/blockchain projects? “We expect 2024 to be the year of POC (proof of concept) and MVP (minimum viable product). After that I expect there will be actual use cases,” Manohar said.
But there are definitely obstacles here, including expansion. Despite recent progress, timely verification of transactions in high-volume decentralized blockchains remains a challenge.
Ben Garfinkel, director of the Center for Artificial Intelligence Governance, said in an oft-cited paper published in 2021, wrote “Established permissionless blockchains, including Ethereum, are too inefficient to run anything other than fairly simple applications.” Even apps that “check who won a chess match” are working Breaking through Ethereum’s current limitations.
Magazine: Breaking into Liberland: Evading guards, decoys and diplomats in inner tubes
Still, Garfinkel acknowledged that if smart contracts “become reliable enough,” they could serve as verification mechanisms for international agreements aimed at governing artificial intelligence systems.
Casper Labs is decidedly more optimistic. “In the race to solve the ‘black box’ challenges of artificial intelligence, blockchain is emerging as the all-in-one solution we’ve been waiting for, incorporating much-needed transparency,” Manohar wrote in the report. Artificial Intelligence Systems The inner workings of a system are essentially invisible to the user, hence the “black box” metaphor.
Hybrid blockchain solutions
However, if blockchain technology cannot even scale, how can it possibly be considered a solution to the “black box” problem of artificial intelligence?
“The solution to the scaling problem is through hybrid blockchains,” Manohar told Cointelegraph. No one today is talking about putting massive datasets on Ethereum or Casper Labs’ own layer-1 blockchain. Casper Labs’ solution involves the use of permissioned (private) blockchains and public (non-permissioned) blockchains.
“People force themselves into this idea that you have to have full permission or you have to be completely open,” Manohar said, explaining further:
“In a hybrid blockchain, you have your own private blockchain. You control it, configure it, and you can make it run at the speed you want because your validator set is limited.”
What about public chains? This is more for version control and record keeping. For example, you might want to register a new version of artificial intelligence on a public chain. “The beauty of this hybrid model is that when you need the immutability of a public chain, you have the option of managing the infrastructure yourself,” Manohar said.
As long as you adequately store the reference on a public blockchain, “you can always be sure that the data cannot be tampered with, because if it is, the hashes won’t match.”
Additionally, anything you want to make auditable, you can put it on a public blockchain because it’s tamper-proof. So, “every time I modify the AI or every time I use a new data set, I send a ping to the public blockchain,” Manohar said.
A big problem with artificial intelligence today is that people don’t know when something will go wrong. But arguably, blockchain provides a way to roll back tapes because they are highly serialized and timestamped.
So if an AI model “starts to show signs of hallucinations or inherent biases, you can simply roll back the AI system to a recent iteration that didn’t have these issues and then diagnose where the problem data came from,” Casper Labs notes on its website.
But others are not convinced that blockchain can solve the “black box” problem of artificial intelligence.
“Describing the ‘transparency’ of blockchain as an antidote to the ‘black box’ problem of artificial intelligence is misleading,” Samir Rawashdeh, associate professor and director of the University of Michigan-Dearborn Artificial Intelligence Research Center, told Cointelegraph.
It does not make the inherent inner workings of machine learning models easier to understand, nor does it clarify how specific outputs are traced back to the original training data.
Rawashdeh said that what Casper Labs is really proposing is a “version control system” that – albeit with some nice features – can be used to “track the development and deployment of artificial intelligence models.”
That said, blockchain can indirectly solve the “black box” challenge by providing an audit trail, Rawashdeh added, helping to ensure the data integrity, provenance and transparency of data sets used to train AI models. But that doesn’t make the actual decision-making process any more explainable.
When machines and humans collude
Looking ahead, concerns arise about general artificial intelligence: Can blockchain help avoid the nightmare scenario of a general artificial intelligence model overturning an election or even starting a war?
“It can actually be of great help,” Manohar replied. Blockchain “will be the best kill switch” for artificial intelligence models, as long as its power “passes through a fully decentralized blockchain.”
That is, the blockchain and its human validators determine whether an AI model gains power. “There is always a kill switch signal, and if all validators agree, they can shut down the network and shut down power access to the AI,” Manohar said, adding:
“It can actually serve as a very effective kill switch for those nightmare scenarios.”
Doubts remain
There are other potential obstacles to the integration of blockchain and artificial intelligence. On the one hand, “there’s a lot of skepticism in the AI community about cryptography,” Dutman said. For many people, cryptocurrencies and blockchain still conjure up images of irreplaceable token scams and other unsavory practices.
That said, when asked if Foresight is seeing more funding proposals for AI/blockchain projects, Duettmann responded: “There’s a lot of movement like that in the space right now.” She sees an average of about five a week. A financing proposal that combines blockchain and artificial intelligence technology. Of course, the institute can only fund a small portion of this, but “it’s certainly getting a lot.”
As for the two communities, “ultimately they have a lot to learn from each other,” she said. In her SmartCon 2023 speech, she noted that the crypto industry is very good at cybersecurity, often employing “red teams,” where teams hunt for inputs that lead to catastrophic behavior. “Let’s extend ‘red teaming’ to machine learning models,” she proposed.
More recognized in China
The combination of artificial intelligence and blockchain technology seems to be particularly popular in China. In Casper Lab’s survey, 68% of Chinese IT respondents agreed that “the convergence of artificial intelligence and blockchain technology has the potential to revolutionize our industry, thereby enhancing data security, transparency and efficiency.” In comparison, The proportion in the United States is 48%, and in Europe only 34%.
Recent: Boosting blockchain adoption by keeping the technology on the backend
Why is it so high in China? Manohar said China has been hostile to cryptocurrencies in recent years but still has a positive attitude toward blockchain technology. Some cities have put land deeds on the blockchain. China views blockchain technology as an effective authentication and tracking mechanism.
In contrast, in the West, “everyone thinks blockchain is just cryptocurrencies,” Manohar asserts. But this educational gap may be closing. In the long run, “everything returns to the mean.”
Is this the killer application of blockchain?
Manohar was asked whether the fusion of artificial intelligence and blockchain would eventually become the “killer app” blockchain has long sought.
“Probably one of them,” he replied. Blockchain tracking governance protocols for supply chain and fintech are also candidates, but these two areas already had decent governance before the advent of blockchain and smart contracts.
In comparison, “there is no existing governance system in the field of artificial intelligence. So there is more room for innovation. So I do think this could be the killer application of blockchain,” he told Cointelegraph.
Svlook
