Co-founder & CTO of Coinscrap Finance
Analyzing world’s first law to regulate AI
According to the press release, MEPs wanted to ensure that companies, especially SMEs, can develop AI solutions without pressure from industry giants. To do this, they will use regulatory sandboxes and real tests, which will be established by national authorities to develop and train the AI before its implementation. The agreement, which came after 38 hours of discussions, has not yet been fully revealed, but there are discrepancies regarding the EU proposal that has been developed during the last two years.
However it may be, Europe has become the first continent to establish clear guidelines for the use of AI. What is this going to mean for our economy? Will it make us more competitive or just the opposite? The EU makes it clear that it intends to encourage investment and innovation, but most of the articles of the AI Law will still take a few years to come into force and many have already spoken out against the rule.
Let’s see in detail what its strengths and weaknesses are:
Pros and cons of the new European AI Law
Pros of this regulation of artificial intelligence
Although they have not provided all the data, the regulation is likely to apply primarily to providers and implementers of AI systems. It has been confirmed that the IA Act will not apply to systems used exclusively for military, defense, innovation or research purposes.
The risk-based approach that the EU has adopted classifies AI systems into four categories based on their use cases: unacceptable risk, high risk, limited risk and minimal/no risk. The law will likely focus on high-risk AI systems. Systems that represent an unacceptable risk will be prohibited.
This last category includes biometric categorization systems that use sensitive characteristics (for example, political, religious, philosophical beliefs, sexual orientation or race), emotion recognition in the workplace and in educational institutions, social scoring based on behavior or personal characteristics or certain predictive policing applications, among others.
Another of the guarantees that accompany the new law refers to chatbots and systems that generate deep fakes, which will be subject to transparency obligations. This includes informing users that they are interacting with AI and tagging artificially generated or manipulated audio, video, text and image content.
Cons of Europe AI Act
As reported by elespañol.com, there are also voices critical of the agreement. The Computer and Communications Industry Association (CCIA Europe) warns that without significant technical improvements in the coming weeks, Europe could miss a unique opportunity regarding the use of AI.
The business association highlights that the text moves away from the approach proposed by the Commission, which prioritized innovation over excessive regulation. They note that certain low-risk AI systems will now be subject to strict requirements without adequate justification, while others will be banned entirely. This could result in a migration of companies and talent, which will seek opportunities to develop in other places.
While companies question the additional restrictions of the new text, the academic community criticizes exactly the opposite. One sector complains about certain legal loopholes that will allow developers to define high risk, while another accuses legislators of leaving it in the hands of the courts to guarantee the fundamental rights of the most vulnerable citizens.
The importance of protected authorship, identity theft and child protection
One of the most worrying cybercrimes that the indiscriminate use of AI brings us is the harassment of minors. Recently we saw how some young people from a small town in Badajoz used a mobile application to “uncover” several of their classmates, and the images began to circulate massively. The mother of one of those affected informed the Police, who quickly identified several people involved.
The bodies had been created with artificial intelligence but the faces were real. This incident generated reflections on the scope of the new tools. Cybercriminals can also impersonate their victims using audios from social media. It is important that the regulation considers the protection of the authorship of these and other content.
With all this in mind, let’s see how the standard can affect our sector:
Traceability of results: financial institutions can feel secure
There is no doubt that banks and insurance companies will find themselves more comfortable working in regulated environments, with traceability of results. The new AIs will also have a point of explainability.
We can suspect something just because we are training it, but now we will be able to obtain certain parameters with its decision. A kind of key to what influences its output.
What does this mean?
Today we do not know what happens “inside” artificial intelligence.
In the middle of the year, CaixaBank announced that they put together a team of more than 100 people focused on the development of applications with generative AI. Without regulation, the reality is that they work in the mud. That is why it is so important to have a regulatory framework that helps all of us, who develop innovative tools, to move in the same direction.
This does not mean that creativity or the possibility of launching original and unique projects is eliminated. On the contrary. Having technological allies is going to be the last resort for the financial sector. At Coinscrap Finance we have been developing modules that incorporate AI for some of the most important banking and insurance companies of the industry for years.
To do this, we collaborate with the University of Vigo and the Center for Technological Development and Innovation (CDTI), which has allowed us to publish several scientific papers. If you want to know more, click here.
About the Autor
Óscar Barba is co-founder and CTO of Coinscrap Finance. He is an expert Scrum Manager with more than 6 years of experience in the collection and semantic analysis of data in the financial sector, classification of bank transactions, deep learning applied to stock market sentiment analysis systems and the measurement of the carbon footprint associated with transactional data.
With extensive experience in the banking and insurance sector, Óscar is finishing his PhD in Information Technology right now. He is an Engineer and Master in Computer Engineering from the University of Vigo and Master in Electronic Commerce from the University of Salamanca. In addition, Scrum Manager and Project Management Certificate from the CNTG, SOA Architecture and Web Services Certificate from the University of Salamanca. He recently obtained the ITIL Fundamentals certification, a recognition of good practices in IT service management.