Artificial intelligence: between progress and ethical dilemma
"We may treat the data we collect and publicly available information to help train our machine learning or artificial intelligence models (...)" or "we use your information to (...) review, improve and develop the Platform, by tracking interactions and usage on your devices, analysing how people use it, and training, testing and improving our technology, such as our machine learning models and algorithms." are you familiar with reading these lines? In case you don't remember, these transcripts are just two examples of the many we could obtain from other platforms, corresponding to the privacy policies of X and TikTok, respectively.
While the rise of ChatGPT has popularised the term artificial intelligence, the idea of creating machines that mimic human thinking is not new. For centuries, philosophers and scientists have dreamt of this possibility, giving rise to works of fiction such as Mary Shelley's Frankenstein or the film 2001: A Space Odyssey by Stanley Kubrick.
The power of artificial intelligence
In today's digital age we are faced with an unprecedented deluge of data
In 1955, the term artificial intelligence (AI) was brought to life by the pioneering mind of John McCarthy. This milestone marked the beginning of a technological revolution that would forever transform the world as we know it. A year later, in 1956, McCarthy, together with a group of visionaries, organised the historic Dartmouth Summer Research Project on Artificial Intelligence. This seminal meeting not only led to the creation of disciplines such as machine learning, deep learning and predictive analytics, but also laid the foundations for prescriptive analytics, a fundamental tool in strategic decision-making.
Beyond these technical innovations, the Dartmouth conference gave birth to an entirely new field of study: data science. This discipline, which draws on AI and statistical analysis, has become a key part of understanding the world around us, extracting valuable information from large datasets and enabling more informed decisions in a variety of fields.
Today, artificial intelligence has gone beyond the realm of science fiction, revolutionising our lives and our ways of working. Since the cases detected in Ayamonte or Almendralejowhere a group of minors were accused of digitally undressing classmates through artificial intelligence, to the new Mango advertising campaign created with generative artificial intelligence, to the recent recreation of a trailer for the popular TV series The Walking Dead with the particularity of setting it in the 1950s, highlight the fact that the advance of artificial intelligence is as important as it is undeniable.
In today's digital age we are faced with an unprecedented deluge of data. Both human and machine activities generate information at a dizzying pace, far outstripping the human capacity to absorb, interpret and make complex decisions based on it. This is where artificial intelligence emerges as the perfect ally to navigate this sea of data and turn it into useful knowledge.
This revolutionary technology is transforming decision-making processes in everything from financial analysis to medical diagnosis.
To illustrate the power of AI, just look at some fascinating examples. Back in 2016, Mark Zuckerberg unveiled Deep Text, a software capable of understanding the context, meaning and sense of messages posted on Facebook with near-human accuracy. On the other hand, in 2024, ChatGPT-4 successfully passed an anti-bot CAPTCHA test by pretending to be a blind person to demonstrate its ability to pass the security test.
These examples highlight the unlimited potential of AI to solve previously unimaginable problems . AI allows us to automate tasks, extract valuable information from large data sets and make better decisions in real time. However, while AI presents immense opportunities to improve our lives and optimise operations, it also raises concerns about its potential misuse. The European Agency for Cybersecurity (ENISA) has identified that cybercriminals are exploiting AI not only to develop malicious code but also to manipulate the behaviours of these AI systems. In addition, the ethical implications of AI cannot be overlooked, particularly with regard to the protection of individual privacy and the responsible management of personal data.
Ethical dilemmas and the regulation of AI
The Artificial Intelligence Regulation (EU) 2024/1689 is based on a risk categorisation approach
The massive collection of personal data, including sensitive data, is crucial to train AI algorithms, improve their performance , and avoid biases or errorswhen processing. However, this practice raises ethical and legal dilemmas around data handling and storage. Data protection is particularly vulnerable when AI algorithms make decisions based on sensitive data without clear explanations of the process. This same case happened to Meta on 14 June. According to the statement issued by Meta due to the demands of the European Data Protection Commission, the company was halting its AI training based on its users' Instagram and Facebook posts across the European Union and the European Economic Area due to a lack of transparency.
In response to these concerns, the General Data Protection Regulation (GDPR) and the Organic Law on Data Protection and Guarantee of Digital Rights (LOPDGDD) have played a key role in protecting privacy in the context of AI until the arrival of the Artificial Intelligence Regulation (EU) 2024/1689 in March 2024 and its official publication in the Official Journal of the European Union (OJEU) on 12 July last, which harmonised rules in the field of artificial intelligence. This Regulation will enter into force on 1 August 2024 and will be applicable progressively as it will not be generally applicable until 2 August 2026, 24 months after its entry into force (which will be 20 days after its publication on 2 August 2026), but some provisions of the Regulation are applicable at different times.
In particular, Chapters I and II shall apply from 2 February 2025; Chapter III, Section 4, Chapter V, Chapter VII, Chapter XII and Article 78 shall apply from 2 August 2025, with the exception of Article 101; and Article 6(1) and the corresponding obligations of this Regulation shall apply from 2 August 2027.
The new law is based on a risk categorisation approach, establishing different levels of restriction depending on the risk potential of each AI system. In the case of unacceptably riskysystems, such as those using biometric categorisation, image extraction for facial recognition, emotion recognition, social scoring or behavioural manipulation, the law provides for a total ban. This measure responds to the need to protect citizens from possible abuses and violations of their fundamental rights.
It is worth noting that prior to the adoption of the European law, ISO 42001:2023 was published in December 2023. This international standard provides a framework for organisations to develop, implement and operate AI systems in a safe, reliable, transparent, ethical and accountable manner. ISO 42001:2023 is applicable to any organisation, regardless of its size, sector or legal form, and includes both public and private entities.
The combination of the new European law and ISO 42001:2023 marks a significant milestone in AI regulation. These frameworks provide companies and organisations with the necessary tools and guidelines to develop and use AI in a responsible manner, ensuring respect for human rights and ethics in its application.
With extensive experience in digital regulatory consulting, Izertis is well positioned to help organisations thoroughly analyse and comply with new regulations, ensuring they harness the power of AI responsibly while navigating the ever-evolving legal landscape.