Blockchain will help clean up and validate the data used in artificial intelligence applications, make predictive capabilities more accessible and allow individual people to treat their data as a commodity, experts told the AiDecentralized conference.
Orgnaized by the Association for Computing Machinery (ACM) and held in Toronto on Tuesday, AiDecentralized brought together researchers and practitioners in the AI space with those working with distributed or replicated ledger technologies on blockchain and cryptocurrencies to explore synergies.
While the two technologies are the easily the most discussed building blocks for innovation in 2018 so far, the areas where they intersect have not always been clear. Depending on the technologies involved, for instance, AI can be used to analyze data and predict future patterns and trends. Blockchain, meanwhile, has served as a foundation for cryptocurrencies such as Bitcoin and Ether but also for a range of applications in the enterprise, such as “smart contracts” which make it easier to exchange items of value.
“AI makes the car driverless — blockchain buys you the car, and upgrades you to a Lamborghini along the way,” joked Lior Messika, who has cofounded an investment and research firm called Eden Block. “We’re not here to disparage blockchain, but it only scratches the surface of what AI will be doing.”
In fact, while many speakers at AiDecentralized criticized the challenges of scaling blockchain, its security risks and other issues, they said it might complement AI in areas of data quality. Dawood Khan, CEO of blockchain certification firm TransformationWorx, pointed to things like the No Fly List, which may have a high rate of “false positives” identifying people as a criminal or terrorist, as an example.
“If that (data) goes into machine learning and decisions start being made with it, that’s a pretty scary world for me,” said Khan, speaking in an AiDecentralized panel discussion. Instead, blockchain could be used to make such data immutable before it gets used by AI.
Paul Lee, founder and CEO of Mind.AI, suggested the rise of “tokenomics” — whereby people are paid via cryptocurrencies — could be a way of widening the talent pool for worthy but difficult to fund AI initiatives.
“We want to democratize AI — and we need a community around the world to contribute,” he said. It might be difficult to hire people, however, when you’re competing with firms like Facebook and Google which can offer extremely high salaries. Paying with tokens, however, means those working on AI projects to end poverty or move humanity forward in some fashion have an alternate kind of payment to offer. “Blockchain incentivizes in ways that no other technologies have done.”
Those incentives can be applied to more than just a developer community, said Susan Oh, CEO of MKR AI, which uses AI to flag false information online. As more activity takes place through digital channels, individuals should have more ownership of their own data and be compensated accordingly.
“The old model is, ‘I push ads off to you,’” she said. ‘Blockchain says, ‘You’re going to help me generate this value, so I need your permission to take a look at this, and reward you with tokens.’”
This notion of “tokenization” may make some business decision-makers uncomfortable, Khan said, and it may take some time to educate enterprises about how cryptocurrencies could help them achieve their objectives with AI.
“There have been tons of proof of concepts,” he said, but many of them have failed to see the light of day. “There’s a lack of ability for the group that’s doing the proof of concept to show business value.”