Artificial intelligence (AI) and blockchain are emerging digital technologies that are capturing the public imagination, but also raising serious concerns.
So it’s worth asking: can AI and blockchain be integrated in a way that benefits humanity?
There are reasons to think so. Back in 2016, Vitalik Buterin wrote that both the cryptoeconomics and AI security communities were “trying to tackle a fundamentally same problem”: regulating complex and smart systems with “unpredictable emergent properties.”
Both depend for their control on essentially ‘dumb’ systems ‘whose properties once created are, after all, inflexible’. For example, once a smart contract is deployed, it cannot be changed. The two communities “need to listen more to each other,” he concluded.
Over the past year, with the rise of ChatGPT and other generative AI tools, there has been growing concern that AI is spiraling out of control. Humans could lose control of autonomous weapons systems in one nightmarish scenario.
So the idea that blockchains and smart contracts can somehow serve as guardrails to prevent AI models from going off course has become increasingly popular.
“Anyone working in crypto has a very clear role to play in getting AGI running smoothly,” said Allison Duettmann, president of the Foresight Institute, at the recent SmartCon 2023 conference. This is especially true given predictions that artificial general intelligence, or AGI, where machines achieve human-level intelligence, could emerge sooner rather than later.
This potential merger of AI and blockchain technology was also on the minds of IT decision makers who participated in a recently released survey commissioned by Casper Labs. Nearly half (48%) of the 608 IT leaders surveyed in the United States, Europe and China agreed that “the integration of AI and blockchain technology has the potential to revolutionize our industry , enabling improved data security, transparency and efficiency.”
Complementary technologies, growing momentum
The basic idea is that blockchains’ immutable, tamper-free ledgers, along with smart contracts, can provide the guardrails for AI implementations, ensuring responsible artificial intelligence. Some believe that a blockchain could even serve as a kind of kill switch for out-of-control AI models.
In the Zogby Analytics survey commissioned by Casper Labs, 71% of IT leaders said they “view blockchain and AI as complementary technologies.” Furthermore, when asked how their organizations currently use blockchain, “working efficiently with AI was the most popular answer (51%).”
Elsewhere, US President Joe Biden issued an executive order on November 1 setting new AI safety and security standards. The order aims to protect the public from a wide range of risks, including dangerous AI-engineered biological materials, AI-enabled fraud and deception.
That order “created a lot of momentum,” Mrinal Manohar, CEO and co-founder of Casper Labs, which has a layer-1 blockchain aimed at enterprises, told Cointelegraph in an interview. AI governance is on the minds of more and more business IT people today.
Does he see more companies launching actual AI/blockchain projects? “We expect 2024 to be a year of major POCs [proofs-of-concept] and MVPs [minimum viable products]. And after that, I expect there will be actual use cases,” Manohar said.
But there are certainly obstacles here, including scaling. Quickly validating transactions in high-volume decentralized blockchains remains a challenge, even though progress has been made recently.
In a widely cited 2021 article, Ben Garfinkel, director of the Center for the Governance of AI, wrote that “established permissionless blockchains, including Ethereum, are too inefficient to run anything other than fairly simple applications.” Even an application “that checks who won a chess game goes against Ethereum’s current limitations.”
Magazine: Breaking into Liberland: evading guards with inner tubes, decoys and diplomats
But if smart contracts “ever become sufficiently reliable,” Garfinkel conceded, they could be useful as verification mechanisms for international agreements aimed at governing AI systems.
Casper Labs is clearly more optimistic. “In the race to solve AI’s ‘black box’ challenge, blockchain is emerging as the all-in-one solution we’ve been waiting for to integrate much-needed transparency,” Manohar wrote in the report. The inner workings of AI systems are basically invisible to users, hence the ‘black box’ analogy.
The hybrid blockchain solution
But how can blockchain technology ever be seen as a solution to AI’s ‘black box’ problem if it can’t even scale?
“The way you address the scale problem is through hybrid blockchains,” Manohar told Cointelegraph. No one today is talking about putting massive data sets on Ethereum or on Casper Labs’ own layer-1 blockchain. Casper Labs’ solution includes the use of both permissioned (private) blockchains and public (unauthorized) blockchains.
“People have forced themselves into this kind of thinking where you have to have full consent or you have to be completely open,” said Manohar, who further explained:
“In a hybrid blockchain you have your own private blockchain that is yours. You manage it, you configure it, and you can run it as fast as you want because you have a limited validator set.”
And the public chain? That’s more for version control and record keeping. For example, maybe you want to register a new version of AI in the public chain. “The great thing about this hybrid model is that you choose when you need immutability from the public chain and that you simply manage your infrastructure yourself,” says Manohar.
As long as you adequately store the reference on the public blockchain, “you can always ensure that the data hasn’t been tampered with, because if it was tampered with, the hashes wouldn’t match.”
Plus, anything you want to make verifiable can be put on the public blockchain because it cannot be tampered with. So “every time I adjust the AI or every time I use a new dataset, I send a ping to the public blockchain,” Manohar said.
A major problem with AI today is that you don’t know when something will go wrong. But blockchains provide a way to rewind the tape, so to speak, because they are highly serialized and time-stamped.
So if an AI model “begins to show signs of hallucinations or inherent biases, you can simply roll the AI system back to a recent iteration that was missing these issues, and then diagnose where the problem data came from,” Casper Labs notes. website.
But others aren’t convinced that a blockchain can solve AI’s “black box” problem.
“It is misleading to describe blockchain’s ‘transparency’ as an antidote to AI’s ‘black box’ problem,” Samir Rawashdeh, associate professor and director of the Dearborn Artificial Intelligence Research Center at the University of Michigan, Dearborn, told IPS. , to Cointelegraph.
It does not make the inherent inner workings of a machine learning model easier to understand, nor does it clarify “how a given output relates back to the original training data.”
What Casper Labs is actually proposing, Rawashdeh suggested, is a “version control system” — albeit with some nice features — that can be used “to track the development and deployment of the AI model.”
That said, a blockchain could indirectly address the “black box” challenge, Rawashdeh added, by providing an audit trail that helps ensure data integrity, provenance and transparency in the data sets used to build AI models. to train. But it doesn’t make the actual decision-making process any better.
When machines team up against humans
Looking ahead, concerns are emerging around artificial general intelligence: could blockchain help avoid those nightmare scenarios where AGI models overturn elections or even prosecute wars?
“It could actually help a lot,” Manohar replied. Blockchain “would be the best kill switch” for an AI model, provided the electrical current “passes through a fully decentralized blockchain.”
That is, the blockchain and its human validators decide whether the AI model is given power or not. “There’s always a ‘kill switch’ signal, where if all the validators agree, they can just shut down the network and cut off the AI’s access to power,” Manohar said, adding:
“It could actually act as an incredibly powerful kill switch for those nightmare scenarios.”
Suspicions remain
There are other potential barriers to this integration of blockchain and AI. For starters, “there’s just a lot of distrust within the AI community around crypto,” Duettmann said. Crypto and blockchain still bring to mind non-fungible token scams and other unsavory behavior for many.
That said, when asked if Foresight was seeing more funding proposals for AI/Blockchain projects, Duettmann responded: “There’s a lot of movement in the space now.” She sees an average of about five financing proposals per week that combine blockchain and AI technology. Of course, the Institute can only finance a fraction of this, but “a lot has certainly been raised.”
As for the two communities, “ultimately they have a lot to learn from each other,” she said. In her SmartCon 2023 talk, she noted that the crypto industry is very good at network security, often using “red teaming,” where teams look for inputs that cause catastrophic behavior. “Let’s extend red teaming to machine learning models,” she suggested.
More acceptance in China
The integration of AI and blockchain technology seems to be viewed particularly positively in China. In the Casper Lab survey, 68% of Chinese IT respondents agreed that “the integration of AI and blockchain technology has the potential to revolutionize our industry, bringing improved data security, transparency and efficiency become possible.” By comparison, that share was 48% in the US and only 34% in Europe.
Recent: Driving blockchain adoption by keeping technology on the back end
Why so high in China? China has been hostile to cryptocurrencies in recent years but remains positive about blockchain technology, Manohar noted. Some municipalities have put land deeds on a blockchain. China considers blockchain technology as an effective certification and tracking mechanism.
By comparison, in the West “everyone thinks blockchain is just cryptocurrency,” Manohar claimed. But this education gap is likely to narrow. In the long run, “everything reverts to the mean.”
Is this killer app from blockchain?
Manohar was asked whether the merger of AI and blockchain could ultimately amount to blockchain’s long-sought ‘killer app’.
“It could be one of them,” he replied. Blockchain’s track-and-trace management protocols for the supply chain and financial technology sectors are also candidates, but these two areas had reasonable governance before blockchains and smart contracts ever appeared.
In comparison, “There is no existing governance system in AI. So there is much more room for innovation. So I really think this could be the killer app of blockchain,” he told Cointelegraph.