Leaders including European Commission President Ursula von der Leyen, former U.S. Secretary of State Hillary Clinton and Danish Prime Minister Mette Frederiksen will this week throw their weight behind a new institute that aims to scrutinize how safe AI applications are for kids.

The Youth AI Safety Institute wants to take on the role of a watchdog and a ratings agency that creates standards to evaluate the safety of AI products for children and scores applications based on how child-friendly they are.

It comes as the impact of social media and other tech on the health of minors has triggered a wave of proposed restriction, with many European countries pushing for bans on social media. Three in four European citizens think governments should set minimum ages for social media, a POLITICO European Pulse survey showed last month.

Von der Leyen has made online child safety a core priority of her second mandate as head of the EU executive.

The institute will be presented in Copenhagen on Tuesday, with von der Leyen scheduled to give a speech. Other big-ticket names backing the institute include former European Commissioner Margrethe Vestager and AI “godfather” Yoshua Bengio.

Experts at the Youth AI Safety Institute will test AI products to inform parents and policymakers on potential harms for children, the organizers told POLITICO. The initiative is run by U.S.-based non-profit Common Sense Media, which already tests movies, series, books and games to provide age ratings.

“The first set of standards and evaluations we will publish will be focused on the most severe harms” such as “suicide and self harm, eating disorders, dangerous advice,” said Bruce Reed, head of AI at Common Sense Media who previously advised former President Joe Biden on AI policy.

Danish politician Vestager, who ran digital policy at the European Commission during its last mandate, said in an interview that “this is really a bold move” to monitor and scrutinize harm to kids from AI “instead of saying it’s too difficult [and resorting to] keeping kids away from AI forever.”

The work to scrutinize products and publish evaluations will be conducted with San Francisco research lab Transluce, non-profit Humane Intelligence and Stanford Brainstorm, the university’s lab for mental health innovation.

Some of the funding behind the institute comes from philanthropists, hedge funds and venture capitalists. Anthropic, Pinterest and OpenAI’s foundation will also be contributing funding.

Asked whether tech firms would have an influence on the work of the institute, Reed stressed that “funders have no say in testing, scoring or outcomes. The Institute will do what Common Sense Media has done for decades: provide tough, honest evaluations of the risks tech products pose to youth.”

The European Commission is considering guidelines for AI companions under its powerful platforms law, the Digital Services Act, EU official Martin Harris-Hess, in charge of minors protection, said at POLITICO’s AI & Tech Week on Wednesday.