This match made in Heaven becomes more evident as we explore both the good and bad aspects of AI-powered technology. I moderated a discussion on Ethical AI & Blockchain at the AI & Blockchain Virtual Expo 2024, on October 31, to delve deeper into this topic, which is also a passion for our Founder. If you missed the event, here is a summary.
Panelists had a wide range of experience – from technical and research, to associations, corporations, to health care. Their diverse perspectives provided fascinating insights into the hot issue of AI ethics, and how we can solve and prevent the major concerns surrounding it. The group included Wei Zhang (Research Director of nChain), Kristopher Klaich (Policy Director of the Digital Chamber of Commerce), Kerstin Eimers, Global Web3 of Deutsche Telekom, and Daniel Pietrzykoski (Technical Strategy & Product Innovation – Blockchain & Program Management of Johnson & Johnson).
The panelists discussed AI ethics and the concerns that surround AI, with real-world examples.
Klaich & Eimers explained that we are still at the experimental stage of these future technologies. This brings up scams and other challenges. Klaich gave the example of AI agents working in the blockchain industry, creating their wallets and acting like influencers. They could pump a token up or offer small amounts of cash to others to promote it. This can be used to manipulate the market and make it difficult to distinguish between a real person and a fake.
Klaich said, “The possibilities are endless and a bit terrifying in this sense.” Eimers cited apps that generate “ideal” AI generated pictures and headshots. These can cause mental health issues in young women, Eimers said. Eimers said, “I think that there are many areas to be monitored. They are amazing on the one hand, but they also bring with them ethical struggles and challenges.”
Zhang revealed that regulation is his biggest challenge. He used an analogy about a criminal using a fast vehicle to escape a scene, but the solution was not to ban cars. It was to ensure that police officers also had fast cars.
He said, “This leads to an overall conflict between technology advancement and regulations and then misuse of technology.”
AI is being misused and we need to create corresponding technology in order to combat the new technology. Zhang said that if we ban technology blindly, he will have police without cars.
Pietrzykoski was having technical problems throughout the panel and was able to comment on this point.
There’s a lot uncharted ground right now and I think many people are trying figure this out. There is a great deal of apprehension, and fear. “Either we clarify what we need to, or something bad will happen, and we will have to find a way to deal,” he said.
How can we handle it before something bad occurs? Here’s where blockchain and education can help.
Klaich spoke about the importance of privacy, ownership, provenance, and how we can hash data to the blockchain in order to trace the data used for the AI models. Zhang, who is involved in research on verifiable AI using blockchain, said that the key to regulation, whether it is led by the government or the industry, is striking a balance of data utilization and privacy.
I’m tackling another angle here, whether we can find a way to resolve the conflict completely between data privacy and data usage. “Of course, this is too good for true, but the seeds are there. This is using blockchain and cryptography,” said he.
It is the idea to prove that the input into the AI model and the execution of AI model as well as the output of AI model can be verified without compromising the privacy of users. We will not be revealing any trade secrets about AI models, but we can be confident that the model is performing as expected.
In this scenario, certification of an AI model’s training on a specific data set is one use case. This proves the model’s unbiasedness. Or, the data owner can collect royalties when an AI model uses their data. This also requires issuing of a certificate.
The blockchain-based micropayments open up new doors for users willing to share personal data.
“With AI, and applications that run AI, or are based upon AI we will be able to choose whether we interact with these models and receive micropayments or compensation for releasing certain pieces our personal information,” Klaich explained. Eimers is passionate about the role of education to prevent misuse of AI. She was happy to elaborate on this topic. “Education is a very important part of the process and goes beyond simple materials or checklists for our customers,” Eimers said.
Eimers shared details about Deutsche Telekom’s awareness campaign, including “A Message from Ella,” which was a deepfake child designed to raise awareness among parents regarding data privacy.
We invest a lot in time and resources to get these messages out there and use our reach as effectively as possible. She said, “I think this is very, very important, and that corporates have to take responsibility and be accountable.”
The issue of who is to blame if AI fails was also raised, a topic that was hotly debated because we are still on uncharted territory. No one has the answer yet. Zhang is confident that we cannot hold technology responsible and says shutting down tech will never be a solution. Remember, the police also need fast cars.
I would say that the first technical step to accountability is to create identities.
In Web3 or any future version of the internet, only one identity protocol exists…In this system, your interactions are auditable and accountable. You can track and verify all history using blockchain. We can protect user privacy by using cryptography and only reveal it if a crime has been committed,” he said.
Zhang said, “We hope that we can have at least a system to help identify the misbehaviors of AI models, and, therefore, to hold their developers or companies accountable.”
For artificial intelligence (AI), to be able to thrive and work within the law, it must integrate an enterprise blockchain system. This will ensure data input quality and ownership. It will also allow it to keep data secure while also guaranteeing its immutability.Read CoinGeek’scoverage of this emerging technology to learn more.Why Enterprise blockchain will be at the core of AI? Watch: Blockchain and AI unlock possibilities