In a significant move to regulate the ever-evolving field of artificial intelligence (AI), the European Union on Thursday adopted the AI Act. This legislation aims to strike a delicate balance between promoting innovation and mitigating potential risks associated with AI.
Passed by an overwhelming majority in the EU parliament, with 523 votes in favour and 46 against, the AI Act introduces stringent regulations for developers of high-risk AI systems. These regulations target advanced AI models like OpenAI's GPT-4 and Google's Gemini, compelling developers to address potential threats such as accidents and cyberattacks.
Transparency lies at the heart of the AI Act, requiring AI developers to disclose the data used to train their systems and comply with EU copyright laws. Moreover, high-risk products, including medical devices and educational systems, face additional obligations to ensure safety.
Non-compliance with these regulations carries hefty penalties, The EU can slap AI providers with fines ranging between $8.2 million and $38.2 million, or 1.5-7% of a company’s global turnover, depending on the size of the violation.
The AI Act also emphasises the importance of consumer awareness when interacting with AI technologies. For instance, deepfake images produced using AI must be clearly labelled, and chatbots must disclose that they are AI-powered during interactions.
Certain AI applications, such as predictive policing and emotion recognition systems, have been outright banned in workplaces, schools, and social settings due to their potential risks. Additionally, the law prohibits police officers from using real-time facial recognition technology, except in cases involving serious crimes like rape or terrorism, and under strict judicial oversight.
While Europe takes the lead in AI regulation, discussions on similar regulations are ongoing worldwide, including in countries like China, the United States, and India. The AI Act marks a significant milestone in global efforts to ensure the responsible development and deployment of AI.
