More than 100 international organisations have called on governments to hold artificial intelligence companies directly responsible for protecting children online, warning that existing regulatory approaches are failing to prevent harm.

The appeal, led by the children’s rights organisation 5Rights Foundation and supported by groups including Amnesty International and Save the Children, was issued ahead of the United Nations’ first Global Dialogue on AI Governance. The coalition argues that governments should require technology companies to prove their AI systems are safe for children before they are made publicly available.

The organisations said artificial intelligence is already affecting children in harmful ways and that current regulations often respond only after damage has occurred.

Leanda Barrington-Leach, executive director of the 5Rights Foundation, said children are not opposed to technological progress but want stronger safeguards in place before products reach the market.

“They aren’t asking us to block AI innovation, but it shouldn’t be a case of cleaning up the mess after harm has happened either,” she said.

The statement comes as legal challenges involving AI companies continue to grow. Firms including Character Technologies and OpenAI have faced lawsuits over allegations that AI-powered chatbots, particularly “companion” bots designed to simulate emotional relationships, can negatively affect children. Some claims also argue that these products have been presented as suitable for younger users without sufficient warnings about potential risks.

The coalition believes governments should focus on the commercial incentives that encourage companies to prioritise rapid product development and user engagement over safety.

Among the recommendations are mandatory safety assessments before AI products are released, financial penalties for companies whose systems violate children’s rights, bans on design features that exploit children’s psychological vulnerabilities, and restrictions on the commercial use of children’s images, voices and biometric data.

The organisations maintain that governments do not need entirely new legislation to introduce these protections. Instead, they argue that countries should enforce commitments already made under international agreements, including the UN Convention on the Rights of the Child and the Global Digital Compact, which outlines principles for protecting human rights in the digital environment.

Barrington-Leach said stronger accountability is needed to change how AI products are developed and marketed.

“As long as companies are rewarded for speed, engagement and data extraction rather than safety, we’ll keep treating the symptoms while the disease becomes endemic,” she said. She added that respecting children’s rights should become a fundamental requirement for technology companies rather than a voluntary commitment.

The United Nations’ inaugural Global Dialogue on AI Governance is expected to bring together government representatives, technology companies, researchers and civil society organisations to discuss international standards for artificial intelligence. Child safety is likely to be one of the key topics as policymakers seek to balance technological innovation with stronger protections for vulnerable users.

France’s startup ecosystem is undergoing a period of rapid change, driven by the rise of artificial intelligence and growing concerns about the country’s political future, according to new figures released by startup incubator Station F as it marks its ninth anniversary.

The Paris-based campus, regarded as the world’s largest startup incubator, has supported more than 9,000 startups since opening in 2017. Companies that emerged from the hub include Hugging Face, Pasqal and Pollen Robotics.

New data from Station F shows artificial intelligence has become a dominant force among resident startups. This year, the incubator launched its F/AI programme, attracting major AI firms, including Mistral AI.

According to the survey, 77 percent of startups said AI tools have reduced their hiring needs. Despite that, 82 percent reported they are either actively recruiting or expect to hire employees in the coming months.

One of the most widely used AI tools among startups at the campus is Anthropic’s Claude model, now used by 90 percent of teams. The result marks a major shift from last year when OpenAI held the leading position among AI providers at the incubator.

AI-related acquisitions have also become increasingly common. Earlier this year, cloud computing startup Koyeb was acquired by Mistral AI. Previous deals included the acquisition of Pollen Robotics by Hugging Face, reflecting continued consolidation within the sector.

The profile of startup founders is also changing. Station F reported that the average founder is now 36.5 years old, compared with 31 in 2018. Around 20 percent hold doctoral degrees, suggesting that technical expertise is playing a larger role in attracting investment and building companies.

The report also found that startup ambitions have shifted. Half of founders expect their businesses to end in acquisitions rather than public listings. Only 9 percent believe their companies will eventually launch an initial public offering, down from 16 percent a year earlier.

International talent remains a key feature of the campus. Roughly one-third of residents come from outside France, with more than 60 nationalities represented. After France, the largest foreign groups come from the United States, Morocco, Germany, the United Kingdom, Algeria and India.

Despite continued growth, political uncertainty has emerged as a significant concern. More than half of founders surveyed said France’s upcoming presidential election is among their biggest worries. Nearly half expressed concern about a possible far-right victory, while about a quarter were worried about a far-left outcome.

Many entrepreneurs cited immigration policy as a major issue, warning that tighter visa rules could make it harder to attract global talent and build internationally competitive businesses.

The findings highlight a startup sector that remains innovative and internationally connected, but increasingly focused on the political environment shaping its future.