Hello from the digital site,
Patrick from Digi-Pal here. 👋👋

Yesterday, the European Parliament presented the “Artificial Intelligence Act: deal on comprehensive rules for trustworthy AI”

• What did it consist of?

• How the the German Bitkom evaluate the act?

• What is my opinion?

• Official Announcement

Ready? Lets go down the rabbit hole 🐇🐇🐇

Subscribe, to stay up to date to the latest announcements

What does the AI Act consist of?

Barry Scanell, Consultant / Laywer at William Fry LLP quotes, what the AI Act implicates:

1. A tiered system based on risk has been established. High-risk AI systems now necessitate a mandatory fundamental rights impact assessment.

2. Regulation of foundation models, following President Biden’s Executive Order approach, will be applicable to models requiring 10^25 flops of compute power during training, essentially targeting the largest language models.

3. Prohibited systems include biometric categorization using sensitive characteristics, untargeted scraping for facial recognition databases, emotion recognition in workplaces and educational settings, social scoring based on personal characteristics, AI manipulating human behavior against free will, and AI exploiting vulnerabilities based on age, disability, or socio-economic status (Scannell, 09.12.23).

4. Transparency requirements are imposed on high-risk AI systems.

5. High-risk AI systems must address biases effectively, ensuring non-discrimination and respect for fundamental rights.

6. Providers of high-risk AI systems must maintain comprehensive documentation demonstrating regulatory compliance, covering programming and training methodologies, datasets used, and oversight measures (Scannell, 09.12.23).

7. Human oversight is mandated for high-risk systems to minimize risks and involve human discretion in AI deployment.

8. Non-compliance can result in substantial fines, ranging from €35 million or 7% of global turnover to €7.5 million or 1.5% of turnover, depending on the infringement and company size (Scannell, 09.12.23).

Businesses engaged in now-prohibited technologies face strategic challenges. Enhanced transparency may conflict with protecting intellectual property, necessitating a balance between disclosure and trade secret maintenance.

Adherence to the documentation requirements poses an administrative burden, potentially affecting the time to market for new AI products.

Incorporating human oversight into high-risk AI systems requires design changes, deployment adjustments, and potential staff training.

The substantial fines for non-compliance represent a significant financial risk, emphasizing the need for legal advice in navigating the new regulatory landscape (Scannell, 09.12.23).

Statement of the Act by Bitkom CEO

According to Bernhard Rohleder, CEO of Bitkom, we have not learn anything by the GDPR-Regulation.

In his opinion he quotes that “The agreement on the AI Act represents a political showcase success at the expense of both the economy and society. The compromise reached last night, particularly in regulating generative AI, overshoots the mark and delves deeply into technology.

The EU is thus imposing a regulatory burden on companies. There is a significant risk that European businesses may struggle to keep pace with rapid technological developments due to impractical regulations. Regrettably, there has been a departure from the previously sought-after application-based and risk-based approach. General Purpose AI Models are now regulated as a technology in themselves, irrespective of specific applications. While the established and generally welcomed instrument of mandatory self-regulation through Codes of Practice is intended to be utilized, it only applies to a small portion of the overall requirements, with the majority rigidly enshrined in the law. Consequently, there is a danger that this crucial field of AI may be constrained by requirements that cannot adapt to emerging technological advancements, a challenge already evident in the negotiations leading up to the AI Act.

Although a fundamental compromise has been reached on this central and highly contentious issue on paper, the substantial challenge lies in translating the overnight agreement into practical rules that form the basis for responsible AI management. The ongoing danger of impeding, rather than facilitating, the application and development of AI in Europe remains.

Experiences with the General Data Protection Regulation also demonstrate that articulating correct goals on paper is insufficient. The challenge lies in practically establishing a genuine balance between risk management and fostering innovation. Europe has the opportunity to take a pioneering role in the ethical and responsible development of AI. However, this necessitates an uncomplicated and well-structured implementation of the legislation. A targeted dialogue among all stakeholders must commence promptly to achieve this. Companies require swift legal certainty and practical support in implementing the AI Act

(Rohleder, 09.12.23)

My opinion

In the fast-evolving landscape of artificial intelligence (AI), recent developments such as the AI Act have and will spark debates and concerns e.g. Bernhard Rohleder presenting a critical viewpoint that resonates with my perspective. As we delve into the intricacies of AI regulation, it becomes evident that finding the right balance is crucial for fostering innovation and maintaining competitiveness.

Key Technological Landscape
Artificial intelligence is the pivotal technology of our era. Its transformative potential will touch every facet of society.

Global Dynamics
We have already witnessed a loss of momentum in embracing groundbreaking technologies. European businesses possibly seem to lean toward American solutions, indicating a concerning trend that may hinder the Europes ability to shape the future of AI.

Regulation as a Double-Edged Sword
While regulation can be a catalyst for positive change and business opportunities, it also raises the question of whether we are tipping the scale towards an excess of regulatory measures. As the title says, having more regulation than competitive, successful AI companies is subjective but prompts a critical evaluation of the regulatory landscape.

The Growth Dilemma
My view aligns with Rohleder's concern that overly stringent regulation can impede growth and stifle the evolution of new possibilities in the AI domain. Achieving a delicate equilibrium between regulation and innovation is paramount for sustainable progress.

Call for Support and Promotion
Rather than adopting an overly regulatory stance, a more effective approach may involve fostering and supporting AI enterprises.

Example: Aleph Alpha's success, securing 500 million in funding without apparent state aid, illustrates the potential of private support. Investing in homegrown talent and incentivizing innovation could be more beneficial in the long run.

Regulation and Competitive Disadvantages
Excessive regulation not only hampers the growth of AI companies but also discourages businesses from incorporating AI models due to compliance concerns. This potential competitive disadvantage could result in substantial economic losses for Germany and the EU.

The Attraction of Global Giants: In the face of stringent regulation, companies may find it more appealing to turn to large American corporations. The allure lies not just in avoiding regulatory hassles (which will be implemented by the big companies) but also in the perceived ease of implementation.

In conclusion, while regulation is essential for responsible AI development, a nuanced and well-balanced approach is imperative. Striking the right equilibrium will not only safeguard against potential risks but also position Europe as a frontrunner in the ethical and responsible advancement of artificial intelligence. The challenge lies in steering clear of overregulation and embracing a supportive environment that encourages innovation without compromising on ethical standards.

We will see how the AI landscape will develop in Europe. Even though I hope for the best, I fear that we will lose the battle.

Official Announcement