My Points:
— In four years, under Ursula von der Leyen, Europe has shifted from simply setting guardrails on AI to now adopting an AI-first approach, no matter the circumstances.
— European taxpayers’ money should be carefully planned.
— Ursula von der Leyen should outline clear priorities for where to invest and where not to allocate resources (human, monetary, hardware), and how to efficiently allocate taxpayers’ funds.
— It’s not just about being AI-first in everything; for instance, many AI startups are closing — how to safeguard a future, how to safeguard the funds, how do we utilize what they’ve created?
— How can we steer the energy of startups in the EU and elsewhere in the right direction? Not every startup will become a billion-dollar success, but every small contribution should count.
— The larger their impact, the more it should matter. How the EU would collaborate with the world to keep its startups safe, and lead to an AI that can matter in the future. How to hire such professionals who can mentor this work.
Here are details of AI acts and AI laws from 2021 to the present under von der Leyen.
April 2021:
— Initial proposal.
— The AI Act
2023:
— A legal framework on AI apps, tools and applications.
— Need for trustworthy AI.
— Defined banned AI
— Regulate artificial intelligence
— Identifying risks
— To scale up the EU AI community.
— Guaranteeing the safety.
— Opens EU supercomputers to European AI startups.
— Responsible development, deployment.
— Development of global guardrails for trustworthy AI.
— AI systems that negatively affect safety or fundamental rights.
— 1 billion euros per year from the Horizon Europe and Digital Europe programmes in AI research and innovation.
— Around 100 companies have already expressed their interest.
— “The ban of AI systems posing unacceptable risks started to apply on 2 February 2025”
2024: I call it the Guadrails Act. The act comes into action.
— The safety and governance requirements, I call it the “guardrails act”.
— AI systems developed or used within the EU should be secure, reliable, and uphold fundamental rights.
— Some systems are considered harmful and hence put to scrutiny before being put into use.
— Some AI systems are considered a threat to people’s and countries’ safety and should be avoided. These include. Care for vulnerable people such as toddlers and children, scoring people based on their use of social media, use of biometrics in public areas, and emotion recognition.
— A lot of high-risk systems were described, from transport to medicine to education to justice and law.
— Some AI is free from obligations of further testing, such as spam filters and AI games. However, AI games, too, can pose some risk in certain cases.
February 2025
— Prohibited AI rules should begin to apply.
— Gives time to apply the rules to Aug, 2025.
— €200 billion mobilization
— €20 billion for AI gigafactories
April 2025:
— AI continent plan
— AI gigafactories
October 2025:
— AI first mindset -in all areas of AI (robotics, healthcare, science applications ..)
— Apply AI applications against previously Guardrails in AI.
— AI Continent Action Plan and the Apply AI and AI in Science Strategies to boost EU incentives in AI with €1 billion funding.
— AI service Desk to help with stakeholders’ queries.
— 27 member states to have similar regulations.
— The 2024 act said some systems are harmful to health and have health concerns; however, Ursula von der Leyen said in this week’s meeting of the EU that she is herself a trained doctor and is amazed by the progress AI has made in medicine. Medicine is not yet fully ready for AI. This is a contradictory statement within a year.
— Interface in Jan 2026 to get on with all member nations of Europe.
— Launch of Apply AI-AI First in October 2025
Thank you for reading.
Subscribe for updates.
References
[3] https://ec.europa.eu/commission/presscorner/detail/en/statement_23_6474
Artificial Intelligence, Computer Science & Mathematics 