Motion for a Resolution by the Artificial Intelligence in a Digital Age (AIDA)

AI-Venture: Artificial Intelligence (AI), with its most recent developments, has already brought numerous advantages in front of issues the global economy is fighting: ageing populations in advanced economies and low productivity in developing economies. From job displacement, to rising inequality and occupational health, AI appears to be menacing the stability of many jobs. How can the EU ensure the stability of its labour market and of its workforce, while also making the most out of what can be better achieved with technologically advanced instruments?

Submitted by: Isilde Cachia (NL), Hendrik Cotterell (NL), Summerjay DuVall (NL), Giorgia Gambellini (IT), Oscar Grimmelt (NL), Odysseus Abraham Kirikopoulos (GR), Nadir Luijten (NL), Hanna Robbens (NL), Kjelt Weyprecht (NL), Matthieu Chiagano (Chairperson, IT)

The European Youth Parliament aims to achieve a balance in the EU’s policies between the benefits and disadvantages of using AI on a broad, daily scale. It appreciates AI enhancing the efficiency and productivity of the workforce and acknowledges that it could bring forth grave challenges such as widespread disinformation, unemployment, and social inequality. Furthermore, we seek to do so by educating the public on the usage of AI and its deriving dangers, as well as by supporting the stakeholders in their fight against job displacement, social disparity, and AI-generated social bias in the workplace. 

The above is decided upon because

• The persistence of misinformation that has come with the recent AI boom leads to:
  1. the inclusion of  data decay¹, low-quality data or incorrect data, since AI is trained with material derived from the internet, 
  2. the generation of inaccurate outcomes, which in turn could lead to harmful consequences such as misleading court cases, 
• The outcomes provided by generative AI lack sources as chat models are being trained to replicate language,
• Access to digital tools continues to be uneven and limited to specific groups, as the digital divide combined with the advent of AI increases inequality 
• Furthermore, the consequences of the above mentioned reason have led to an indiscriminate loss of jobs,
• The implementation of AI in the working environment is likely to increase inequality across society in different groups, such as:
  1. lower-skilled workers compared to higher-skilled workers, since lower-skilled jobs can be replaced more easily, taking wealth into account,
  2. the old and the young due to the differing levels of knowledge and power in decision-making processes regarding AI in the workforce in the future,
• Developing countries, not having access to the same resources and opportunities as developed countries, face reduced benefits from the usage of AI,
• AI systems have been called inhumane, unethical, immoral by some experts, leading to:
  1. an overall negative impact on humanity of AI-made decisions,
  2. a wider spread of bias, such as black prisoners being actively discriminated by AI risk assessment programme, further worsened by a general and unquestioning trust towards AI-made decisions,
  3. AI being ethical or morally capable of doing certain jobs since they need human engagement, such as therapy,
• People can jailbreak²  or misuse AI for malicious purposes that undermine public safety such as manipulating them into generating harmful and objectionable content, such as the possibility to utilise AI to be provided bomb-making instructions,
• The implementation of AI in the workspace is prognosed to affect 40% of all jobs by 2030, possibly resulting in a decline in mental and physical health of the employees, 
• Research on the safety of AI is already scarce, and there is a reluctance to implement a globally adopted safety net, causing other Member States to fall behind in the development of AI technologies,
•  AI that is trained on copyrighted material without giving appropriate credit to the copyright owners, is leading to a loss of the value of the original intellectual property and the original creators themselves,
• Job instability and job loss caused by the implementation of AI can affect the financial well-being of workers and other citizens too.

To that end, the European Youth Parliament

1. Appeals to the European AI Office to introduce a team of trained specialists to screen the existence of widespread malicious prompts on the web and manually correct the database that AI bases answers on;
2. Suggests AI companies³ to enforce a limit on AI sourcing its materials to only reliable factual sites, such as scholarly articles, scientific publications and professional news sources;
3. Encourages Member States to cooperate with the Directorate-General for Education, Youth, Sport and Culture (DG EAC) to educate citizens on the dangers of misinformation and foster curiosity and critical thinking, thus accustoming the public to fact-check the information that AI presents by enriching their academic programme;
4. Invites the European Cybercrime Centre (EC3) to establish a wide database with known jailbreak prompts and keep up with the newly discovered prompts, and to further enforce the database’s usage on AI models to prevent them from being jailbroken;
5. Instructs the European AI Office to forbid publicly sharing jailbreak prompts and their usage, when it’s not being used for research purposes;
6. Calls upon the European Commission to further update the AI Act to:
   • require that trustworthy AI⁴ is to be programmed to  present sources and licence the information that it uses to create generative AI, following the example of journalistic tailored AI models such as Perplexity AI 
   • introduce a clause within the AI Act making it mandatory for AI to give the appropriate credit to the creators of the material that it uses when producing an answer;
7. Asks the Directorate-General for Communication Networks, Content and Technology (DG CONNECT) to address the digital divide by allocating funds for Member States, supporting them in increasing digital accessibility⁵ in less digitally developed regions;
8. Invites Member States to bridge the generational gap by educating elderly people through factsheets and open workshops on how to use AI properly;
9. Suggests national Ministries of Education include mandatory classes about AI in their school curricula to endorse citizens to use AI in a responsible way , learning of its dangers and developing their critical thinking process;
10. Requests DG Connect to study and implement a list of professions that always require human presence and declare to what extent AI is allowed to be involved in them; 
11. Appreciates the European AI Office’s effort to continue enforcing the AI Act⁶, ensuring that AI is unbiased, neutral and non-racist;
12. Invites Member States to support local enterprises in being alert of a possible unemployment crisis by backing them in keeping a necessary minimum amount of employees;
13. Calls upon Member States to create a mandated free re-education programme on the use of AI in the workplace for employees whose job is directly replaced with AI;
14. Welcomes Member States to exchange information on AI usage with developing countries through workshops, information exchange programs and digital learning courses by implementing institutions such as the Ministry of State for Artificial Intelligence of the UAE in their national institutional frameworks;
15. Encourages DG CONNECT to address copyright infringements by implementing guidelines for the introduction of a digital helpdesk⁷ on AI websites to indicate when creators’ work has been copied without giving the right credit.

Lastly, the European Youth Parliament recommends participants to

• Take part in panel debates similar to the panel debate at National Conference Utrecht 2024 aimed at improving media literacy, and actively look for more similar opportunities through local and national institution’s social media pages, such as Utrecht University’s Panels and Workshops online section;
• Educate themselves on the responsible use of AI models and get a better understanding of their functioning through free courses such as Codecademy;
• Use alternative AI systems, based on scientific papers or studies, which give credit to original creators, such as Consensus or Scite;
• Challenge themselves into trying to recognise AI-generated speech through a small 2-minute conversation on the online game Human or Not.

1. Data decay is the gradual deterioration of data quality over time. ↩︎
2. Jailbreaking is a way to break the ethical safeguards of AI models. ↩︎
3. AI companies are companies that specialise in integrating artificial intelligence solutions to help other companies improve both productivity and efficiency, and remain competitive in their industry. ↩︎
4. Trustworthy AI is a lawful, ethical and robust AI model. ↩︎
5. Digital accessibility means that everyone can use the possibilities offered by the internet, computers and smartphones. ↩︎
6. The AI Act is the first-ever legal framework on AI, which addresses the risks of AI and positions Europe to play a leading role globally. ↩︎
7. A digital helpdesk is a contact centre that serves employees or customers to provides assistance and information. ↩︎