A recent horrible example of sham reporting has brought the worst-case scenario of combining unethical professional behavior plus AI to the forefront. A magazine editor was fired over fabricating a front-page AI-generated 'interview' with seven-time F1 champion, Michael Schumacher, who suffered a tragic injury ten years ago and has not been in the public since that time.
Misinformation and deep fakes are already rampant problems. Now, this dangerous situation is being supercharged by the capabilities of publicly available generative AI tools like ChatGPT. I can't help but imagine the societal harm that could be unleashed by one bad actor using this capability to fake comments from world leaders.
What can we do to rein in this problem?
I often look for solutions in the combination of other powerful technologies with AI. In this instance, I see tremendous potential for Blockchain+AI.
Using blockchain and AI together to address deep-faked interviews would be like placing a lock and a security camera on a valuable item. The lock (blockchain) would provide a secure and tamper-proof way of storing the item, while the security camera (AI) helped to verify the authenticity of the item and ensure that it has not been tampered with by bad actors. Just as a lock and security camera can protect a valuable item from theft or damage, blockchain and AI could protect the integrity of interviews and prevent them from being manipulated or misrepresented.
Blockchain and AI could be used together to create a decentralized and tamper-proof system for authenticating interviews with politicians, experts, and celebrities. I certainly don't profess to be a Blockchain expert, so I look forward to hearing from those who know far more on that topic than I do. But here are some starting thoughts for why I believe this solution merits focus.
Here's how it could work:
Record the interview: The interview could be recorded using a camera or audio recording device equipped with AI capabilities, such as speech recognition and natural language processing. This would help to ensure that the recording is accurate and complete.
Store the interview on the blockchain: Once the interview has been recorded, it could be stored on a blockchain network, which would provide a decentralized and tamper-proof way of storing the data. This would help to ensure that the interview cannot be modified or tampered with without leaving a trace.
Use smart contracts: Smart contracts could be used to automate the process of verifying the authenticity of the interview. For example, the smart contract could be programmed to verify the identity of the interviewee and interviewer, as well as the date, time, and location of the interview.
Use AI for verification: AI could be used to verify the authenticity of the interview by analyzing the speech patterns and language used by the interviewee and comparing it to their known speech patterns. This could help to ensure that the interviewee is not an impostor or that the interview has not been doctored using deep fake technology.
Provide public access: Once the interview has been verified and authenticated, it could be made available to the public on the blockchain network. This would provide a transparent and decentralized platform for sharing interviews with politicians, experts, or celebrities while also ensuring their authenticity.
In this way, using blockchain and AI together could provide a powerful mechanism for authenticating interviews and ensuring their accuracy and authenticity.
Such an approach would also need to:
Address privacy concerns: My proposed approach should ensure that the privacy of the interviewee is protected, especially if they are a private individual rather than a public figure. The system should also be designed to comply with data protection laws.
Consider scalability: My proposed approach should be designed for scalability, as it may need to handle a large number of interviews. This could involve using a scalable blockchain network or implementing a sharding mechanism.
Ensure security: The system should be designed with security in mind, to prevent unauthorized access or modifications to the interview data. This could involve implementing strong encryption and access controls.
Address technological limitations: While AI can be used to verify the authenticity of interviews, it is flawed and has limitations. These limitations should be taken into account when designing the system, and alternative verification methods should be considered.
Ensure user adoption: The success of such an approach would depend on user adoption, so it should be designed to be user-friendly and easy to use. It may also be necessary to provide incentives for users to participate, such as offering rewards for contributing verified interviews to the network. And ultimately, regulators would need to enforce adherence.
The recent incident of a magazine editor fabricating an AI-generated celebrity interview highlights the urgent need for a solution to combat the rampant problem of misinformation and deepfakes. The combination of unethical professional behavior and AI capabilities can have devastating consequences, especially if used to fake comments from world leaders.
The potential harm is alarming, but there is hope in the combination of blockchain+AI. By creating a decentralized and tamper-proof system for authenticating interviews with politicians, experts, and celebrities, we can ensure their accuracy and authenticity. While this approach requires addressing privacy concerns, scalability, security, technological limitations, and user adoption, it has the potential to revolutionize public discourse and restore trust in media. Regulators must enforce adherence, but this is a step towards safeguarding the integrity of public discourse.