The new generation of chatbots has taken the web by storm, but they are not without problems. Their ability to delirate and spread misinformation is truly chilling, and doubts about the ethical development of the models behind the scenes are multiplying rapidly. Mozilla has not only recognized this situation but has decided to move forward with a long-term solution: Mozilla.ai, a new start-up aimed at building an open and transparent artificial intelligence ecosystem.
Anyone who has had a conversation with ChatGPT or Bing Chat in recent weeks has probably noticed two things: the authority these chatbots convey when responding, and their extraordinary ability to invent things or reproduce garbage. Although there are topics the chatbots refuse to discuss (for different reasons), the most likely outcome is that sooner or later they'll conjure up something completely false or riddled with errors.
From fake news to copyright conflicts, many arrows point to a problem with the models themselves and their training processes. Not to go too far, OpenAI is being accused of abandoning the Open in its name, and limiting the amount of public technical details about GPT-4. For the people at Mozilla, that strategy is wrong, and with the help of researcher Moez Draief they have decided to create a new start-up called Mozilla.ai.
Mozilla.ai: Can a Trustworthy AI Be Built?
If you go to the official Mozilla.ai page, you'll find nothing more than a brief introduction and a contact section, but that's more than enough: instead of launching a chatbot or its own model, Mozilla.ai's plan is to develop an open source, decentralized, and trustworthy artificial intelligence ecosystem, backed by tools that will provide security during the creation of recommendation systems and generative platforms.
The announcement on the Mozilla blog goes a bit further, mentioning that artificial intelligence can enrich people's lives as long as the interests of users are not lost sight of, and parameters such as transparency and responsibility are prioritized. However, Mozilla is not offering mere words: the initial investment in the company is 30 million dollars. We'll have to wait a few months to learn the next steps of Mozilla.ai, but my suggestion is that they shouldn't take too long. The "ChatGPT Era" has already arrived, and the chatbot isn't asking permission, so to speak...
Official site: Click here
