The wide-spread adoption of generative AI by New Zealand organisations could leave them damned by association in the case of crisis, new research shows.
The new research, published in Business Horizons by Professor Dan Laufer, head of AUT’s School of Communication Studies, and Professor Yijing Wang of Erasmus University Rotterdam’s Department of Media and Communication, identifies five main types of organisational crises caused by generative AI and offers practical guidance for organisations to proactively protect themselves from reputational damage.
Laufer says they have seen a lot of organisations in New Zealand enthusiastically adopt generative AI for many purposes.
“Often they are using generative AI as a unique selling point, to promote their innovative credentials. However, there are now several examples of spillover crises caused by various applications of generative AI, and organisations should learn from these crises and proactively prepare for future crisis communications.”
Professor Dan Laufer.
The research identifies five main types of generative AI crisis:
- Authenticity/Integrity – related to content trustworthiness and erosion of credibility when using Gen AI
- Labour displacement – employment risk associated with the use of Gen AI
- Technical failure – related to concerns about the perceived reliability of Gen AI tools
- Data security and privacy concerns – data management when using Gen AI
- Discrimination and bias – concerns about systemic bias and structural inequality
The authors say a ‘spillover crisis’ occurs when a crisis affecting one organisation or individual creates the concern, uncertainty or perception or harm for another organisation or individual.
Laufer says perception is crucial when considering spillover crises.
“We’ve gathered examples from each category, and it’s clear that the public assumes all organisations in a category behave in similar ways and that they are also guilty by association.
“For example, Sport Illustrated’s use of fake author profiles for GenAI-generated articles caused sector-wide scrutiny of editorial authenticity in journalism.
“The best way organisations can inoculate themselves against these spillover crises is to have a clear stance on their use of GenAI, and the ways they are using it, and being willing to disclose whether their organisation uses GenAI tools that are involved in the crisis or not.”
Laufer and Wang’s research shows that clear, unequivocal statements help distance organisations from potential crisis spillover.
“The most successful inoculations against potential spillover crises are when organisations act quickly and decisively to distance themselves. However, in order to do that, they must already have clearly defined the GenAI tools they are using, for what purpose,” they say.
Read Laufer and Wang’s article in Business Horizons
