The stock on AI and disinformation: green light from the European Parliament for the rules on the use of AI with to the AiAct
Further regulations will be incorporated at European level into the new overall regulatory framework on artificial intelligence about online content, disinformation, and protection of social users: the “AiAct”, which involves in the forefront the new regulations on ChatGPT and the use of deepfakes.
This is the first global regulation designed to protect users from the possible risks of digital technology and has been approved on the 14thof June in the Strasbourg plenary. It was authorized by 499 votes for, 28 votes against and 93 abstentions.
Already on 5th of June, from the European Commission the Vice-President Věra Jourová clashed with Twitter. The social giant, in fact, had decided not only not to release the report on the anti-disinformation measures of its platform, but also decided to withdraw from the EU Code of Conduct against online disinformation. The code was created back in 2018 and currently sees the membership of 44 large companies, to name a few: Google, Meta, Tiktok and Microsoft. The company of Elon Musk had chosen to fight, forcing the European Commission to take note of it by ensuring compliance with EU rules with greater rigour and urgency.
Moreover, the vice-president added that if Twitter “wants to operate and do business in the European market, it must comply with the Digital Services Act”, the Digital Services Act (DSA) and the Digital Market Act (DMA) published in the Official Journal of the European Communities on 27th October.
The main objectives include the protection of the fundamental rights of all users of digital services, the achievement of equal conditions to promote innovation, growth, and competitiveness both within the EU and globally and the protection of users from fake news and disinformation.
The first steps towards the implementation of the legal text began the last 15th May and saw MEPs meet in joint meeting. The central theme was facial recognition in public places, which also raised some concerns.
Here are some of the themes of the text:
1.Use of cameras for facial recognition: the use of biometric technologies at AI will be banned in real time inside public places; therefore, the proposal of the EPP that wanted instead, reduce its use only in cases of missing children and cases that would compromise national security has been rejected. Although, as far as border management is concerned, the subject may still be open.
- Inclusion of prohibitions on possible invasive and discriminatory uses of AI
- Biometric categorisation systems based on sensitive characteristics (such as race, ethnicity, religion, or political orientation)
- Insurance policy (to analyse the habits and behaviours of customers trying to avoid risks)
- Emotional recognition systems used by law enforcement in border management, educational institutions, and workplaces.
In addition, the text envisages categorising those applications that are of greatest risk, considering those AI systems that could affect safety, health, fundamental rights and the environment. At risk, are also considered those systems of AI that allow to influence the political outcomes of elections in social media platforms, the most influential ones with over 45 million users.
As for Chat GPT, the AI system is required to comply with the conditions of transparency, stating what content was generated by AI and which not, so that users can distinguish deep fakes from real images, including possible copyright.
Regarding the use of AI, the Vice-President of the Commission said: “I have said many times that we have the task of protecting freedom of speech, but when it comes to artificial intelligence, I do not see any right for machines to freedom of speech”. Words that make us think about the constant debate between AI and people in an era that has become increasingly digital. We wonder, therefore, if with the AiAct the distances between us and the machines will be shortened, if the latter can act as a support in distinguishing what is real from what is not.
Ultimately, as for fake news, From 25th August this year, the fight against disinformation will become a legal obligation in the EU with the effective entry into force of the European Digital Services Act, which requires platforms to take further measures of transparency in content.