Incentivai Uses AI Simulations To Check For Blockchain Hacking Exploits
Cryptocurrency projects can succeed or fail based on a single blockchain exploit. As blockchain technology becomes more prevalent, the consequences of blockchain exploits become more severe. A startup called Incentivai recently came out of stealth to solve that problem.
Incentivai uses artificial intelligence simulations that test for security holes and other potential blockchain-based exploits. Incentivai doesn’t just look for problems within the code: it looks for exploits that greedy or illogical humans could exploit in a blockchain.
Incentivai plans to offer its AI-powered service to blockchain developers before their platform goes live.
One of the biggest problems with blockchain exploits is that once the platform is released in the wild, it can be hard to solve the exploits. Once a smart contract is up and running, for example, and coins start to get involved, it can be difficult to implement fixes. There’s no rewind button in the blockchain world. Blockchain’s immutability and irreversibility is one of its biggest selling points, but also one of the main downsides.
“There are many ways to check the code of a smart contract, but there’s no way to make sure the economy you’ve created works as expected,” says Incentivai’s solo founder Piotr Grudzień in a statement to TechCrunch from earlier this weekend.
“I came up with the idea to build a simulation with machine learning agents that behave like humans so you can look into the future and see what your system is likely to behave like.”
Incentivai came out of stealth on Friday, August 17. Next week, the company will graduate from the Y Combinator accelerator program. Incentivai already has multiple customers.
Incentivai offers two core services: customers can pay Incentivai to audit their project and produce a report. Or, customers can host the AI simulation tool like a software-as-a-service.
As Incentivai explains it, either “you do it” or “we do it”:
- Incentivai Does It: “We use a customised version of the Incentivai simulation tool to prepare a report on the incentive structure of your system.”
- The Customer Does It: “We give you access to a hosted, customised version of the Incentivai simulation tool. You conduct the testing yourself, with our support.”
The first deployments of blockchains checked by Incentivai will be released over the next few months. However, Incentivai has already released multiple case studies to prove the value of its ecosystem.
How Does Incentivai Work?
Using artificial intelligence to identify and resolve blockchain exploits sounds good – but it’s obviously not easy to implement. How does Incentivai solve not just tech problems within blockchains – but human problems?
Incentivai’s founder, Piotr Grudzień, describes the system as similar to “what DeepMind does with AlphaGo”. Incentivai’s AI tests different strategies and measures how those strategies affect the blockchain. Using multiple AI-powered “agents”, Incentivai can monitor how different users can be expected to use the blockchain.
Piotr built his AI development skills while earning a master’s degree at Cambridge before working on natural language processing research for Microsoft.
Here’s the basic process for how Incentivai works:
First, a developer writes the smart contracts they want to test for a product – say, for selling insurance on the blockchain
- Incentivai tells its AI agents what to optimize for, then lays out all the possible actions they could take; these agents can have different identities – like a clueless user fumbling around or a malicious hacker trying to grab as much money as possible
- Incentivai then tweaks its agents to make them more or less averse to risk, more or less technologically advanced, and more or less caring about whether they disrupt the blockchain being tested
- Incentivai then monitors the agents and gets insights about how to change the blockchain system
Using this process, Incentivai might discover that uneven token distribution can lead to pump and dump schemes. Incentivai would recommend that the developer should more evenly divide tokens and give fewer tokens to early users.
Or, if the blockchain is designed to sell insurance, then Incentivai might find that it’s prone to fraudulent claims. The blockchain might require users to vote on which claims are approved, for example, but the company needs to increase its bond price to remove the incentive for voters to take bribes from fraudsters.
By implementing these and other recommendations before launch, Incentivai can help companies avoid disastrous consequences in the future.
There’s no rewind button in the blockchain world. Blockchain is immutable and irreversible, which means any exploits are amplified. Incentivai aims to help blockchain startups avoid exploits before launch using AI-powered agents. Stay tuned for more information about Incentivai. The company just came out of stealth mode last week and is graduating from the Y Combinator accelerator next week.