No ChatGPT if you work for NWO

When distributing research funding, reviewers and collaborators should not use AI programmes such as ChatGPT and Perplexity. This is stated in a new guideline from research financer NWO.

Last year, the Netherlands Organisation for Scientific Research (NWO) already drew a line in the sand: juries, referees and committee members were not allowed to use generative AI for the time being. This ban is now permanent.

Risks

According to this directive, what applicants do is up to them. According to NWO, they are free to use AI as long as they are aware of the risks: for example, fabricated footnotes, ingrained prejudices and plagiarism. Above all, be transparent, is the advice.

But for grant application reviewers, the use of ChatGPT and similar programmes is absolutely out of the question. Just uploading an application would violate confidentiality, NWO believes. You also don’t know if the review is diligent if you leave it to AI.

Approved

Employees of the NWO organisation (lawyers, policy officers and so on) are only allowed to use approved AI programmes, so no ChatGPT, Gemini or Midjourney.

Moreover, they may only use even those approved AI programmes in a limited way. For example, they are not allowed to contribute to draft versions of justifications or defences in appeal proceedings.

As a matter of fact, hardly any AI programmes have been approved yet. According to the NWO website, there is only one: ‘Currently, NWO staff are allowed to use the DeepL translation application.’ (HOP, BB)

Comments are closed.