The European Union's regulation of artificial intelligence was the focus of the 50^th edition of Painel Telebrasil 2024, a major platform of discussion involving representatives of telecommunications companies and government agencies, organized by CONEXIS Brasil Digital and Teletime News. Representatives of European companies operating in Brazil took part in the debate, which was moderated by Daniela Martins of CONEXIS and attended by around 140 people, including parliamentarians, CEOs and representatives of major telecommunications companies.

Marian Schuegraf, European Union Ambassador to Brazil, highlighted that the global AI market is expected to reach €257 billion this year, an increase of €74 billion from last year. In addition, more than 100 million people worldwide are expected to be working with AI by 2025. "AI has significant potential to bring social benefits and economic growth, but certain systems may pose risks to the security of users, including physical security and fundamental rights," she said.

Explaining the need for regulation, the Ambassador emphasized that widely used models could create systemic risks, leading to legal challenges and slower adoption of the technology due to lack of trust. The European Union's legislative measures aim to safeguard fundamental rights and protect democracy. "Our challenge now is to turn this legislation into reality," she noted.

This transition still has a way to go. Maria Buzdugan, First Counsellor at the European Union Delegation to Brazil, explained that following entry into force of the new law on 1 August this year, preparations have started for the implementation phase. The law will be applied as from 2 August 2026, with some exceptions which include the bans on uses of AI considered of unacceptable risk, which should start to apply in February 2025, the obligations for general purpose AI to start in August 2025 and the obligations regarding high-risk uses to start in August 2027.

"The European Commission is working on publishing guidelines for high-risk classification, as well as practical examples of high-risk uses before February 2026," she said, adding that the European standardisation organisations CEN and CENELEC are working to define standards regarding high-risk AI use system, expected to be published in April next year. The Commission is also holding public forums to assist with the drafting of a code of conduct for general purpose AI models, to be published in 2025. “The new rules do support innovation. We believe good regulation can, in fact, promote innovation by enhancing users’ trust in AI applications, which in turn will foster AI uptake,” she highlighted.

Striking a balance

As they push forward with their own AI initiatives, companies operating in Europe are seeking to understand the new regulation. Paulo Bernardocki, Director of Solutions and Technologies at Ericsson, expressed the company's expectation to benefit greatly from AI technology. He highlighted the importance of keeping up with technological developments in order to remain globally competitive. "Our concern is to have the freedom to work with the flexibility required at this moment in time, while preserving values such as user privacy. The challenge is to balance these two elements," he added.

Tiago Machado, Director of Institutional Relations at Vivo/Telefonica, pointed out that the European Union's move was crucial in establishing the concept of risk-based regulation, defining what is unacceptable and promoting best practices. "One of the key discussions is how to separate AI applications that are focused on operational efficiencies from those applications that might require additional care due to risk" he noted.

Paulo Pontes, Director of Ultrasound Latin America at Siemens Healthineers, stressed the importance of maintaining the momentum of innovation. He mentioned that in healthcare, AI can help with early diagnosis in imaging exams such as ultrasound and specific treatments such as radiotherapy. "In healthcare, AI can be seen as a co-pilot that helps physicians make more accurate and faster diagnoses, impacting healthcare access and costs," he said.

Fabio de Freitas, Vice President of ICT and Digital Innovation for South America at Stellantis, highlighted the company’s careful attention to this issue. He noted that customers are demanding innovation in products and services, and that Brazil could look to European practices for inspiration. “Brazil has the potential and talented professionals and can adapt to different scenarios. Balance, flexibility and agility are essential. Committees (in Congress) need to understand this dynamic and the industry is ready for dialogue,” he said.

Paulo Pontes, from Siemens, insisted on the importance of safety and transparency in the process, but warned against over-regulation that could stifle innovation. "Ultimately, we will benefit in healthcare. The key is to balance safety and transparency," he argued.

Tiago Machado, from Vivo, cited ongoing legislative discussions in the Brazilian Senate as an example of progress through dialogue. "We are discussing the balance between ethical and responsible use of AI while maintaining technological innovation capabilities," he said, noting that some points are still being discussed and adjusted. He called for clarity in the law regarding high-risk applications.

He also noted the importance of defining the role of sectoral regulators. "It would be a setback to have a vision that confuses the regulatory scope for a specific sector or creates complex and burdensome regulation. Anatel (National Agency of Telecommunications), for example, understands the importance of using AI in network management," he highlighted.

Paulo Bernardocki, from Ericsson, reiterated that trust is essential and that preserving and protecting users is a company commitment. "What concerns us is that the AI market is developing rapidly. It is crucial to have conditions that allow the technology to develop within the European Union," he concluded.