Mozilla launches $30m startup for trustworthy AI

Initial focus will be on tools that make generative AI safer and more transparent and people-centric recommendation systems

Mozilla, the non-profit behind Firefox, has announced a startup dedicated to providing trustworthy AI.

The move comes as interest in AI reaches fever pitch, thanks in a large part to the success of OpenAI's ChatGPT in bringing AI to the masses. OpenAI, in which Microsoft has a substantial stake, has been joined by Google, Meta, China's Baidu in releasing generative AI products and models, and the field is growing exponentially with many possibilities for science, healthcare, education and the environment.

However, the resources and data required to create large machine learning models means that power is concentrated in a handful of gigantic commercial corporations. A recent research paper by scientists from MIT and Virginia Tech found that large companies are hoovering up skilled researchers too, thus consolidating their grip, with engineers leaving smaller firms to work for them.

"Industry is becoming more influential in academic publications, cutting-edge models and key benchmarks," the researchers note. "And although these industry investments will benefit consumers, the accompanying research dominance should be a worry for policy-makers around the world because it means that public interest alternatives for important AI tools may become increasingly scarce."

There's a danger, in other words, that large scale global applications of AI such as tackling climate change, land use for biodiversity or tropical diseases will be de-prioritised in favour of consumer doodads that earn the big bucks.

The increased competition also means less transparency. OpenAI, which began as a non-profit research organisation, used to share its models and results, but its monicker has become something of a misnomer. The technical report it put out to accompany the release of GPT-4 includes no details about the training set, the architecture of the model or its performance.

In a blog post, executive director of Mozilla Foundation, Mark Surman, writes: "This new wave of AI has generated excitement, but also significant apprehension. We aren't just wondering what's possible and how can people benefit? We're also wondering what could go wrong, and how can we address it? Two decades of social media, smartphones and their consequences have made us leery. "

The company says it is launching Mozilla:AI with initial funding of $30 million as a startup with transparency and openness as core principles "to make developing trustworthy AI apps and products easy". The startup will be headed by Moez Draef, a professor at the LSE who was formerly global chief scientist at Capgemini Invent, and aims to attract thousands of "like-minded founders, developers, scientists, product managers and builders" to its cause of building AI outside big tech and academia.

Its initial focus will be "tools that make generative AI safer and more transparent," and "people-centric recommendation systems that don't misinform or undermine our well-being," Surman writes.