🚦🤖 9 PRINCIPLES FOR REGULATING AI 🤖🚦
Chatgpt has > 1bn users. The AI revolution is here.
So what can regulators do? Here are our first 9️⃣ principles in a thread. Feedback welcome! /1
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628490928570373
1️⃣👷♀️AI SAFETY FIRST👷♀️
Creators of general purpose AI, like ChatGPT, should only be able to operate under a license.
Unsafe, untested, general purpose models should not be allowed on the European market.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628500365713408
2️⃣ 👩💻AI EDUCATION FOR KIDS AND POLITICIANS 👩💻
Every school kid should be taught the basics behind AI, including coding, statistics and ethics.
The same should apply to politicians and regulators, who should receive the same basic training to inform their policy making.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628507449933827
3️⃣🏷️ AI AUTHENTICITY LABEL🏷️
All content that is not fully human-generated has to be marked as such by law.
Labels like “AI generated” or “AI-assisted” content should be developed and enforced, especially for social media to ensure authentic public debates.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628514219589632
4️⃣💶YOUR DATA, YOUR MONEY 💶
Creators should be remunerated, personal data only used with specific consent, providers need to have their data sources ready, antitrust regulators should have sufficient powers and companies should pay reasonable wages for data-labeling.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628521899261955
6️⃣ 🏥OPEN AI FOR GOOD🚋
Where algorithms and training data sets have been developed using public money, they should be available for free.
A Dutch hospital should for example be able to use the cancer detection algorithm developed with Bulgarian public money.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628535828594691
7️⃣ 🚀START-UP FRIENDLY ECOSYSTEM🚀
Across the EU, world-class universities, an attractive European labour migration system, sufficient and secure server capacity should be built.
Our regulatory framework should be easy to understand for everyone building an AI-based start-up.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628544544317447
8️⃣ 👮SMART ENFORCEMENT👮
We cannot yet foresee all needs for regulatory responses to rapid developments in AI.
Principles of AI regulation should be set by legislators, but executive powers should be given freedom to monitor, react to & stop concrete cases of risk or harm.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628551871868928
09 GLOBAL COORDINATION
Let’s share our AI approach via international cooperation.
And we should protect ourselves against spillover from unethical and risky AI applications from elsewhere.
Let’s cooperate globally with our partners to establish global AI principles!
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628558997929989
Great - this was our first draft list of principles. Thanks for checking them out! 🕵️♂️
But now it’s up to you - what do you think we should do to regulate AI? Do you agree with us? 🙋♀️
Let me know! 🙏🏼👇🏽✍️❓
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628562391171073
5️⃣📍SMART & LOCAL AI-TAX📍
AI will likely replace many of the jobs and disrupt the decentralized value generation in our economies. We need to re-localize how societies participate in the value generation of AI.
Every payment made for an AI service should be taxed locally.
🐦🔗: https://n.respublicae.eu/d_boeselager/status/1656628529033797632