The hipster effect can be seen as a form of negative feedback that stabilizes the diversity of opinions and behaviours in a society. People + AI could improve efficiency but also reinforce existing biases. Use AI to improve diversity by considering people's unique perspectives and experiences.
As AI-generated work becomes more commonplace, AI-homogeneity means everything would be more similar, and difficult to tell apart.
Jonathan Touboul, a mathematician at Brandeis University in Massachusetts, has studied how the transmission of information through society influences the behavior of people within it.
His conclusion is that, in a vast range of scenarios, the hipster population always undergoes a phase transition in which members become synchronized with each other in opposing the mainstream. In other words, the hipster effect is the inevitable outcome of the behaviour of large numbers of people.
Similar effects occur among investors, in other areas of the social sciences and with artificial intelligence.
Chatbots are like mirrors held up to society — they reflect back what they see. Algorithmic auditing is a very new process of looking at an algorithm and examining it for bias. It can help to make sure that algorithms agree with what people think is right and do not hurt or treat anyone unfairly. Algorithms need to be checked because they are used more and more to decide things that matter for people in areas like health, education, law, money, and social media.
Another way to limit bias in AI is to set boundaries for how a particular type of system will be used. Salesforce, for example, is limiting its generative AI to answering questions of the sort that a sales or support staffer would need to ask.