French charities must get past ‘shadow AI’ tension to realise potential

While most professionals at French nonprofits are using artificial intelligence tools in their work, this is often ‘shadow AI’, warns a new report by EFA member France générosités.

And the organisation has urged the sector to address this internal mismatch, and think carefully about what AI could really achieve.

Nearly nine out of 10 (86%) of the 375 people surveyed for L’usage de l’IA au sein des OSBL say they use AI at work — with 29% using it every day, and another 47% using it several times a week.

But just over half (55%) of those AI users are doing so outside of frameworks or structures put in place by their organisation – making it so-called ‘shadow AI’.

The most common use of AI is for editing (88% of respondents are doing this), followed by research and monitoring 69%) and communication (36%), while only 15% are using it in their fundraising strategy.

The report finds that senior staff and directors more often use tools recommended by their organisation, while more junior staff are more likely to be shadow AI users. It also says that leaders consistently see their organisation as more sophisticated in its AI use than their teams do.

Nearly two in three respondents (66%) use ChatGPT, while around two in five use Microsoft Copilot (42%) and a fifth (22%) use Google Gemini. Users of AI employ on average 2.5 different AI applications.

Despite organisation’s lack of consistency and the tension between leader’s views and those of their colleagues, France générosités says that the report is inherently promising, commenting:

“Respondents envision real added value from AI in areas that are still relatively untapped: data collection, search, and management. The perceived potential is significant. However, moving from envision to practice requires structured support, identified resources, and appropriate governance.

Also this month, Fundraising Europe reports that UK fundraisers are keen to use AI, but need better guidance, case studies and funding in order to make it a more effective part of their work.