Lisbon, July 21, 2024 (Lusa) - The rules on prohibited artificial intelligence (AI) practices will apply from February, under the European regulation (AI Act), legal expert Sara Rocha, TMC associate at CMS Portugal, told Lusa in an interview.
The European regulation, which was published in the Official Journal of the European Union (EU) on 12 July, comes into force 20 days after publication and must be implemented by August 2026.
AI ‘is the topic of the decade and, as such, lacked regulation’, and with its publication it applies directly to all member states, says the TMC - Technology, Media and Communications associate at CMS Portugal.
‘What the regulation does provide for are extended periods of time for some of its measures to come into force. After its publication, the regulation will enter its transitional period, with most of its rules being applicable within 24 months from the date of entry into force, with other time frames for the application of certain matters,’ she pointed out.
Of particular note is ‘the case of the rules on prohibited artificial intelligence practices, which will apply within six months,’ i.e. from February, she added.
‘The main reason for adopting these implementation phases is, given the complexity of some of the measures, to give companies time to adapt,’ explains Sara Rocha.
The AI Act ‘is based on a risk-based system, providing for some uses of AI that are considered unacceptable and therefore prohibited, such as subliminal behavioural manipulation, exploitation of individuals considered vulnerable, social scoring by the government or private individuals that leads to discrimination, and remote biometric identification in real time in public spaces by police authorities, with certain exceptions,’ she explained.
Asked about her expectations regarding its implementation, she said that its ‘main objective is to promote the development and use of AI in the EU, as well as to ensure a high level of protection, while at the same time trying to stimulate the use of AI and the development of the sector’ and also provides for ‘concrete obligations in order to protect citizens’.
If there was a time when ‘the advantages/disadvantages of cloning or using technology were debated, the debate now oscillates between the advantages of using AI - freeing us from some work, helping to respond to some diseases, speeding up many responses, automating processes - and the disadvantages or fears that arise with its use - increasing social inequalities and discrimination, being uncontrollable systems or making it impossible to distinguish man from machine’.
In response to this, the European legislator approved a regulation with ‘68 definitions, 113 articles, 13 annexes and 180 recitals’, which is ‘a piece of legislation with a very complex structure that can only be read by a niche that is already familiar with this type of language, understanding most of these rules is not simple and excessive complexity is not always the best option for innovation - which is one of the main flagships’.
In this context, the lawyer said that the expectation will be to ‘continue to develop AI systems that are useful to the population, without losing speed in their development, but to say that these tools respect citizens and the legal system in which they are inserted, while being aware of these difficulties’.
ALU/ADB // ADB.
Lusa