The EU institutions started on Wednesday their discussions on the forthcoming Artificial Intelligence (AI) Act but the European Commission has already published internally guidelines for staff on the use of online available generative AI tools.
The European Parliament adopted its negotiating position on the AI Act on with an overwhelming majority - 499 votes in favour, 28 against and 93 abstentions. The rules would ensure that AI developed and used in Europe is fully in line with EU rights and values including human oversight, safety, privacy, transparency, non-discrimination and social and environmental wellbeing.
The Commission proposed already in April 2021 new rules and actions aiming to “turn Europe into the global hub for trustworthy”. AI is much more complex and riskier than other IT tools.
A key point in the Commission’s proposal was therefore a risk-based approach, where AI systems that are considered a clear threat to the safety, livelihoods and rights of people will be banned. High-risk AI systems will be subject to strict obligations before they can be put on the market. AI systems such as chatbots are considered to have limited risks but users should be aware that they are interacting with a machine so they can take an informed decision to continue or step back.
As reported after the parliamentary vote, the examination of the Commission’s proposal took time. While AI can save lives by advancing medical diagnoses, they can also be exploited by authoritarian regimes to carry out mass surveillance of citizens. The debates were further delayed in recent months by controversy over the dangers of generative AI which is capable of creating texts or images.
Generative AI systems like ChatGPT are based on machine learning algorithms trained on vast quantities of data. According to the EU co-legislators they would have to comply with transparency requirements and ensure safeguards against generating illegal content. Detailed summaries of the copyrighted data used for their training would also have to be made publicly available.
These AI systems can easily provide answers to all possible questions and summarize the content of vast amounts of information. In the future they may make experts in different areas redundant as the digitization of routine tasks changed the labour market. But already media abounds of reports of misleading or directly false replies giving by the new AI tools.
Civil society activists are aware about the risks of inbuild bias in the AI systems and argue that the algorithms should take into account for example human rights and animal welfare concerns. While safeguarding the systems against risks, they can also be used to promote citizen participation.
A Commission spokesperson told The Brussels Times that the Commission published its guidelines on 31 May. These guidelines aim at assisting staff in understanding the risks and limitations that online available tools can bring and support them in appropriate usage. The guidelines are internal to the Commission and were communicated to all staff.
| The guidelines list 5 concrete rules: · Staff must never share any information that is not already in the public domain, nor personal data, with an online available generative AI model. · Staff should always critically assess any response produced by an online available generative AI model for potential biases and factually inaccurate information. · Staff should always critically assess whether the outputs of an online available generative AI model are not violating intellectual property rights, in particular copyright of third parties. · Staff shall never directly replicate the output of a generative AI model in public documents, such as the creation of Commission texts, notably legally binding ones. · Staff should never rely on online available generative AI models for critical and time-sensitive processes. |
Any official can download ChatGPT. With the use of ChatGPT, there is no problem to improve speeches. A Commission official said that he mostly used it for drafting letters and speeches. “You just have to ask the programme to improve the text and you’ll get a perfect speech in flawless language.” Whether the efficiency was achieved at the expense at new ideas and deeper insights he did not tell.
M. Apelblat
The Brussels Times
