When I do keynotes, I have regularly shown two slides, which might also be of interest for people that read my blog. As the AI Act timeline should be clear by now, this post will focus on four suggestions of how to develop a competitive edge.
Slide #1 shows the key implementation dates:
Slide #2 underlines what stakeholders should IMO do now to get a competitive edge:
▪️Standardisation: Article 40 will likely stand for the easiest and cheapest way for providers to prove that their high-risk AI system are compliant with the #AIAct. In order to allow European and national standardisation bodies to come up with adequate and balanced standards, it is however crucial to involve civil society, academics as well as SMEs and start-ups. Please provide your technical expertise based on your AI systems! Some contacts if you seek help:
➡️ CEN and CENELEC: via Sebastian.Hallensleben@vde.com or https://www.cencenelec.eu/get-involved
➡️ DIN Deutsches Institut für Normung e. V.: via filiz.elmas@din.de or https://www.din.de/de/mitwirken
➡️ European Commission: via tatjana.evas@ec.europa.eu
➡️ Federal Ministry for Economic Affairs and Climate Action: via buero-vic5@bmwk.bund.de
▪️Secondary legislation: The legal drafting of the #AIAct is rather vague and many details need to be further specified. Facing huge time pressure, the new #AIOffice will rely heavily on external contributions. In 2024, stakeholders should thereby concentrate in particular on 1️⃣ Codes of Practice (Article 56), 2️⃣ Implementing Acts on GPAI evaluations (Article 92(6)), 3️⃣ Guidelines (Article 96(1b/f)) on Article 5 and on the AI definition. Send your feedback (i.e. concrete use cases of AI systems that don’t pose a high risk, technologies that are not AI) to:
➡️ European Commission: via Kilian.Gross@ec.europa.eu + gabriele.mazzini@ec.europa.eu (CNECT-A2@ec.europa.eu in :CC)
▪️Regulatory Sandboxes: As a rather new element of EU digital legislation, the details of RSB remain unclear and differences among Member States seem likely. Yet, they could become very useful in particular for SMEs and startups if designed and financed adequately by Member States. Give it a try, improve your AI systems in a safe space and benefit together with the public authorities from mutual regulatory learning. Some contacts if you have further question:
➡️ European Commission: via Kilian.Gross@ec.europa.eu + gabriele.mazzini@ec.europa.eu (CNECT-A2@ec.europa.eu in :CC)
➡️ Federal Ministry for Economic Affairs and Climate Action: via buero-ic5@bmwk.bund.de
▪️Enforcement bodies: The #AIAct governance system will only work if it’s well equipped with the right talent. Staffers working in those public entities - on EU or national level - need to have the technical expertise to be capable to engage with AI developers. Help the Commission and Member States to identify and motivate AI experts to join them. Please forward the following contacts:
➡️ AI Office: check the job opportunities via https://digital- strategy.ec.europa.eu/en/news/job-opportunities-european-ai-office or send an email to CNECT-AIOFFICE-RECRUITMENT@ec.europa.eu for general inquires.
➡️ Scientific Panel: via CNECT-A2@ec.europa.eu and via Member States (i.e. buero-vib@bmwk.bund.de)
Last but not least, please let Axel Voss and me know (via kai.zenner@ep.europa.eu) if you face problems concerning any of those four proposed actions. We will document your feedback and try to provide support in case there are reoccurring problems.