OpenAI, the developer behind the ChatGPT, no longer plans to transparently document its progress in the field of artificial intelligence in the future. In contrast, the renowned company Mozilla has created a start-up focused on making AI trustworthy, transparent and secure.
There has recently been a lot of criticism of the ChatGPT developer OpenAI. Because the company has decided, contrary to its original principles and its name, to no longer conduct research and development on artificial intelligence openly and transparently. Above all, many researchers and other developers expressed regret about this – also because the construction of the successful language models of the GPT series would not have been possible without open research that could be used by everyone, such as on the so-called Transformer architecture. Mozilla Corporation, which is best known for the development of the Firefox browser and is wholly owned by the non-profit Mozilla Foundation, now wants to take on the role that OpenAI doesn’t want to fill.
“We’re only three months into 2023 and it’s already clear what will be one of the biggest stories of the year: AI,” read the announcement. “AI has captured public attention like Netscape did in 1994 and the iPhone in 2007.” In particular, tools such as Stable Diffusion and GPT-4 would transform the Internet, communications and society in previously unexpected and unimaginable ways. And AI-based platforms like YouTube and TikTok are becoming more and more powerful. The challenge, according to Mozilla, is therefore to research how a trusting future can be created with artificial intelligence.
“We don’t see this happening at the big tech and cloud companies with the most power and influence,” Mozilla said. With a start-up named Mozilla.ai, a “trustworthy and independent open source AI ecosystem” is now to be built. This should make it easier to build AI products that can be trusted. Mozilla.ai should develop itself, but also collaborate with “people who share our vision”. Therefore, the development of tools that make generative artificial intelligence and AI-based recommendation systems “safer and more transparent” should be started.
Mozilla.ai: Artificial intelligence rivel to chatGPT
The company, which was initially equipped with a budget of 30 million US dollars, is to be managed by Moez Draief, who previously researched intelligent systems and quantum computers at Imperial College London and Huawei’s Noah’s Ark Lab in Paris.The subject of artificial intelligence is not entirely new for Mozilla. With the Mozilla Common Voice project, the non-profit had already made a data set for training human voices available free of charge to enable speech recognition and speech synthesis for independent companies and free projects. And as early as 2020, the Mozilla Foundation published the White Paper Creating Trustworthy AI, a frequently cited contribution to the debate on the secure design of AI systems. This includes, among other things, the dangers of unregulated artificial intelligence, the possibilities of misuse and the risk of monopolizing AI systems.
New chance for Firefox?
As Mozilla boss Mitchell Baker said in the Decoder podcast in February, Mozilla sees not only a danger in artificial intelligence, but also a very big opportunity – also for Mozilla itself.
Firefox ‘s shrinking market share. “It’s an opportunity. I think the market for browsers could change,” Baker said.The Firefox browser could be upgraded in the future with AI tools that give users new possibilities. “The question is how much [AI does it need] and what does it have to do?” Baker said. “Do I think there will be disruption? It’s a lot more likely now than it was six months ago.”