Public interest supported by Google The AI ​​partnership launched with $ 400 million + for the construction of the open ecosystem

Make room for another partnership on AI: AI CURRENTwhich is presented as a “public interest” initiative focused on the promotion and development of the direction of artificial intelligence in socially beneficial directions, has been announced to the French AI Summit AI Monday – with a first $ 400 million in donors’ promises and a plan to draw $ 2.5 billion over the next five years.

Such figures may seem a small beer with regard to AI investment, with the The fresh French president of the trumpet of a private support package worth around 112 billion dollars (who himself turns pale next to American investments aimed at accelerating technology). But the partnership is not focused on calculation so that its donors think that such modest sums can always make a difference in key areas where AI could make an essential difference to advance the public interest – whether in AI for Health care or support for climatic objectives.

Under the higher focus on “the environmental environment for the public interest IA”, the initiative has a number of objectives set out – including the thrust to expand access to the sets of high quality public and private data for IA formation; Management of infrastructure and open source tools to increase the transparency and safety of the AI; and support for the development of systems to measure the social and environmental impact of AI.

Its founder, Martin Tisné, said that the objective was to create a financial vehicle “to provide a northern star for public funding for critical efforts”, such as – for example – bring AI to fight against cancers or To offer treatments for a long time to make a lot of confection.

“I think what’s going on is that you have a date bottle of data to come in artificial intelligence, because we lack road with data on the web, and here, what we need , it is to really unlock innovations in the way of making data accessible and available, “he told Techcrunch.

In open source, he declared that the objective was to support the construction of the ecosystem by directing investments in order to guarantee that open source tools “are as transparent as owner tools”.

With regard to the responsibility of the AI, the partnership hopes to “unify the field” – by working on membership in Audit Standards of AI systems which are responsible for the merit of having “a deep implication by different populations and communities that focus on the problems we want [AI to help with]».

“For understandable reasons, there is a lot of concentration on the huge [AI] Investments. It’s different, ”he also told us. “Our goal here is around the public interest. We focus on small models. We do not optimize for act [artificial general intelligence]… We are looking for smaller models where you need specific data sets for high value.

“For example, on Parkinson’s disease – there was an incredible standardization of data sets, highlighted by the Michael J Fox Foundation – as we are looking at really specific things to make a difference in people’s lives.”

Europe and the southern world chip in

The initiative is supported by a mixture of public and private actors – with governments such as France, Germany, Chile, Kenya, Morocco and Nigeria among the nine countries listed as a launch partners (Note: The United States is not Participating, governments across Asia are not either, so the effort of AI is along the way by political decision -makers in Europe and in the world. The other countries listed are Finland, Slovenia and Switzerland).

That said, also listed in PRs as “main partners” are (United States) the giants of Google and Salesforce technology.

On the private sector side, Tisné said that the partnership wanted to work with industry research laboratories doing cutting -edge work (as in life sciences); With technological companies that have a distinct positioning compared to the dominant current, as because of how they and / or their customers use open source, and with other large companies who are users and buyers of buyers open source products; And with startups that push the envelope on the opening.

The other main partners of the current AI who have been appointed to launch are the French government (which is under the projectors of the governance of AI this week because it hosts the AI ​​action summit in Paris) ; With several philanthropic donors – namely: the Ford Foundation; John D. and Catherine T. MacArthur Foundation, and a collaborative – the latter being an organization of lobby of the governance policy AI supported by Omidyar, which is in the billionaire Pierre Omidyar and the network of his wife PAM.

The CEO of collaborative AI, Tisné, is also the founder of the current AI. Discussing the gap The current donors of the AI ​​seek to fill, he argued that there is a gap for a public-private financing vehicle which can gain momentum behind the efforts to generate a development of AI in lines of public interest.

“It is not a lack of public interest projects in AI, it is a massive fragmentation in the field and how we can simply work on a much larger scale,” he told Techcrunch During an appeal, claiming that the initiative aims to conduct “public and private funding on the scale of critical public interest projects on AI”.

“Some of which already exist,” he continued, saying that it was “a question of really bringing them together, but to focus-focusing on the effort so that we can help develop the next Alphafold“, A reference to the pioneering system of the AI ​​of Google Deepmind to predict with precision the structures of the proteins inside the human body.

“The private sector rightly focuses on private interests and works on a large scale – and the fact with, you know, calculate investments in the order of tens of billions per quarter – we are therefore trying to understand what it is is that we could do to really make a difference, “said Tisné, noting:” Alphafold was developed on the basis of a set of public data, including, but without limiting ourselves to the database protein. Thus, a big goal of us will therefore be on data from this point of view. »»

Public interest in Ecosystem Support

Efforts for broad access to health data could, for example, focus on supporting technology development preserving confidentiality to allow more patients to share their data for AI research, he suggested.

“There is not another partnership set up which is really designed to bring together the whole field and to bring together public funding on a large scale,” he said.

Changing the Tisné effort said that the current AI work covers three avenues: first providing financial support to the sector in the form of direct financial contributions. He will also seek to play an incubation role – aimed, for example, to support research work to feed AI innovations. Third, it will operate on “the alignment of funding so that various donors can work jointly on the basis of goals and shared objectives”. Here, his supporters therefore hope to bring together various public interests for supporting AI and amplify the impact.

The partnership will deploy approximately half of its funds given in the form of subsidies. The other half is set for aligned financing efforts – “around the opening, around data and responsibility”, which, according to Tisné, will include “really specific goals and objectives” by program (which are still at define).

“It is not a game of policy or regulation. It’s really a building game, “he added.

AI’s current public relations include a larger list of “supporters” and “champions” for the initiative – citing a letter of open support of a duration of the SPAN Industry, in particular Arthur Mensch, co -founder and CEO of the model of great French language (LLM (LLM) Maker Mistral; Serial entrepreneurs and investors Brent Hoberman and Reid Hoffman; Clement Delangue, CEO of the IA company, hugging the face; and Fiji Simo, member of the Council of ‘Administration of Openai and also CEO and President of Instacart, among others.

In the letter they write: “To reach the best of AI, the company must be in charge.”

“In practice, this means to ensure that high -value data sets are accessible in preserving confidentiality and safe means, encouraging the development of smaller open AI models that meet the needs of people and are more respectful of the environment, and the scale of open source AI to improve to improve to improve and more environmentally friendly, and the scaling of the Open-source to improve to improve to improve and more transparency, security and accessibility for all “, they continue, adding:” The future of AI should belong to all of us. “”

The current AI plans to unveil new supporters and donors in the coming months, Tisné saying that they are particularly eager to work with the Gates Foundation, given the emphasis on the keys as health care.

If you are curious to know their choice of name (AI Current), he said that they were shooting for something that is the field here and now-that is to say AI of the current generation, not things Science fiction that could descend the pipe – as well, like wanting to play on ideas of electric current and even diversity (think of the current of a river and all the life that can browse it).

“It is important to have more diversity in the field of AI,” he said, adding: “We have spent a lot of time in AI to worry about a really distant future and What could happen … so it’s really an effort to focus on opportunities and damage today.