Decentralized AI Offers New Hope for User Data Security

    Published on:

    One of the biggest risks associated with the increased use of AI chatbots like ChatGPT is the emergence of new threats to user data. But some companies are starting to build decentralized AI systems in hopes of making it more difficult to leak personal data.

    AI startup Elna is one of them. The India-based company allows people to create customized AI chatbots on the blockchain while making them “decentralized, transparent” and free to use.

    Also read: Will decentralized digital identity curb AI threats in Web3 and the Metaverse?

    Blockchain security

    “Elna prioritizes user data protection through the use of Internet Computer Protocol (ICP) canisters. [or smart contracts]It is a decentralized container that is directly controlled by users’ wallets,” Elna co-founder and CEO Arun PM told MetaNews.

    “This framework ensures complete ownership, transparency and control of data by users,” he said.

    ICP is a blockchain network (the same kind of technology that underpins Bitcoin) that aims to improve the efficiency and speed of decentralized data storage.

    A major concern with centralized, closed AI platforms such as OpenAI (which seems to have started life as a nonprofit organization dedicated to building open source models) and Google is the vulnerability of personal data. In such settings, data is typically stored on a company's servers, and individual control is often limited.

    PM Alam said that the use of smart contracts in artificial intelligence can solve this problem by protecting privacy and user data such as transaction information and chat history through decentralized storage. In this way, people will be able to control their own information, he added.

    Arun said Elna does not “store or use” user data. He says the model is “essentially consistent” with his EU's strict General Data Protection Regulation (GDPR) and other global privacy laws. Instead, users store their data using Smart Her Contract-based wallets.

    “By leveraging blockchain technology, Elna ensures the integrity and privacy of user data is maintained,” Arun said, adding that the platform collects only the data it needs to operate optimally. .

    “We employ best practices in data minimization and provide our users with tools to effectively manage their data,” he said.

    Source: Elna AI

    What is Elna AI?

    Released in August 2023, Elna describes itself as a “community-driven decentralized AI agent creation platform.” Unlike other AI chatbots where the operator selects the training data, Elna's users can upload their own data to the network and train AI models focused on topics and applications of interest.

    Once that is complete, Elna will deploy an artificial intelligence assistant or AI agent to the ICP blockchain, allowing users to continuously expand the AI’s knowledge by adding new data.

    Elna says the company “Canister Smart Contract” Responsible for training and deploying AI models. It also utilizes what is known as a “vector database” to store information on-chain.

    Users use wallets to log into the system and use ELNA tokens for governance, while Elixir is a utility token and both are built on Internet computer protocols.

    Decentralized AI brings new hope for user data security
    Avatar created on Elna. Source: Elna AI

    ChatGPT data leak

    Big tech companies have failed to protect user data in the past. Centralized entities like Facebook have been proven to play fast and loose with user data. Social media networks exposed private emails, phone numbers, messages, photos, and videos.

    Data itself has become a salable commodity, with the top 50 AI tools being accessed more than 24 billion times last year, according to Writerbuddy, and the stakes are higher than ever. Providing proprietary datasets that artificial intelligence companies collect can be a very lucrative business when that data falls into the wrong hands.

    In fact, in March 2023, a bug was revealed that leaked thousands of chat titles, the first message of a new conversation, and payment information from ChatGPT Plus users. OpenAI I apologized Although there were no issues with leaks, there has been increased scrutiny on how it protects user data.

    As MetaNews previously reported, ChatGPT was temporarily banned in Italy for this very reason. Regulators were concerned about privacy violations as defined by the GDPR.

    University of North Carolina researchers found that centralized large-scale language models such as ChatGPT and Google's Gemini continue to leak sensitive data, including financial records, even after being removed by chatbot creators. I also discovered

    Decentralized AI: Rethinking the wheel

    It’s impossible to tamper with data stored on blockchain, but decentralized AI platforms like Elna don’t come without their own challenges, despite promising a future where users control their data. It does not mean.

    “The main challenge with decentralized AI is the lack of the underlying infrastructure and frameworks present in Web2, and Web3 requires us to rethink the AI ​​wheels,” said Arun PM, CEO of Elna. states.

    He also believes that existing global regulatory frameworks are “not sufficiently equipped to address the unique challenges and privacy concerns posed by centralized AI systems.” Arun said regulators often lag behind major advances in technology, making it difficult to enforce timely data and privacy protection laws.

    But what if Elna's decentralized AI model suffers a security breach, leading to the loss of user data? Is that possible? Prime Minister Arun said such a breach could only occur at the user-created Chabot level, not at Elna's infrastructure.

    “The decentralized nature of the platform enables accurate tracking and attribution of liability,” he elaborated. “This model ensures accountability is maintained by allowing violations to be identified and appropriately addressed.”


    Leave a Reply

    Please enter your comment!
    Please enter your name here