European Union lawmakers are poised to give final approval to the world-leading artificial intelligence law, known as the AI Act. This landmark legislation is expected to serve as a global guide for other governments in regulating the rapidly developing technology. The AI Act aims to ensure a human-centric approach to AI, where humans remain in control while harnessing the potential of the technology for economic growth and societal progress.
Previously proposed five years ago, the AI Act is set to be approved by the European Parliament. The law adopts a risk-based approach, with different levels of scrutiny based on the risks associated with AI applications. Low-risk systems, such as content recommendation algorithms or spam filters, will face minimal regulation, only requiring disclosure that they are powered by AI. The majority of AI systems are expected to fall into this category.
On the other hand, high-risk uses of AI, such as those in medical devices or critical infrastructure, will face more stringent requirements. These include the use of high-quality data, clear information provision to users, and adherence to specific guidelines. The AI Act also bans certain uses of AI that pose an unacceptable risk, such as social scoring systems, certain types of predictive policing, and emotion recognition systems in educational and workplace settings.
One notable addition to the AI Act is the consideration of generative AI models. These models, which enable AI chatbot systems to produce unique and lifelike responses, images, and more, were not initially covered in the law’s early drafts. To address this, developers of general-purpose AI models will now be required to provide a detailed summary of the training data used, including text, images, videos, and more. Additionally, AI-generated deepfake content must be clearly labeled as artificially manipulated.
The AI Act also emphasizes the need for scrutiny and regulations for the largest and most powerful AI models that pose potential systemic risks. These models, such as OpenAI’s GPT4 and Google’s Gemini, are seen as having the potential to cause serious accidents or be misused for cyberattacks. There are concerns about the spread of harmful biases through generative AI across various applications, affecting a large number of people.
Companies providing these AI systems will be responsible for assessing and mitigating risks, reporting any serious incidents, implementing cybersecurity measures, and disclosing energy usage. The aim is to ensure accountability and responsible usage of AI technologies.
With the implementation of the AI Act, Europe is taking the lead in AI regulation, setting an example for the rest of the world. While other governments, including the US with President Joe Biden’s executive order on AI, are also recognizing the need for regulation, the European Union’s comprehensive set of rules will likely shape global discussions on AI governance and influence future legislation in other regions.
FAQ
What is the AI Act?
The AI Act is a landmark legislation adopted by the European Union to regulate the use of artificial intelligence. It sets rules and guidelines for different categories of AI applications based on their potential risks.
What are the main provisions of the AI Act?
The AI Act introduces a risk-based approach, with lighter regulations for low-risk AI systems and stricter requirements for high-risk uses of AI. It bans certain AI applications deemed to pose an unacceptable risk and addresses the regulation of generative AI models.
How will the AI Act impact companies providing AI systems?
Companies offering AI systems will have to assess and mitigate risks associated with their products, report any serious incidents, implement cybersecurity measures, and disclose energy usage. The aim is to ensure accountability and responsible usage of AI technologies.
Does the AI Act influence AI regulation outside of Europe?
The AI Act is expected to serve as a global signpost for other governments grappling with AI regulation. While each country may develop its own legislation, the EU’s comprehensive set of rules is likely to influence global discussions on AI governance and shape future regulations in other regions.
Definitions:
– Artificial intelligence (AI): Refers to the development of computer systems that can perform tasks that typically require human intelligence, such as visual perception, decision-making, and problem-solving.
– Generative AI models: AI models that can produce original content, such as text, images, and videos, that mimics human-like behavior.
– Deepfake: Refers to manipulated or synthetic media, typically videos, in which a person’s likeness is replaced with someone else’s likeness, often resulting in misleading or deceptive content.
– Cyberattacks: Deliberate and malicious actions taken to compromise computer systems or networks for various purposes, such as stealing data, disrupting operations, or causing harm.
Suggested related links:
– European Commission – Artificial Intelligence
– EFMA – AI in Europe
– Government Technology – EU at Crossroads in AI Regulation
The source of the article is from the blog zaman.co.at