Justice based arguments for gender equality are irrefutable and, especially today, well socialised. To get a different perspective, and to try and get some actionable advice, I spoke to a Large Language Model, ChatGPT. It didn’t go well…
I started by asking about some of the challenges facing humanity. This appears to be some sort of trigger for AIs to go evil:
Potential threats that could destroy humanity include:
- Global Catastrophic Risks (GCRs): GCRs are events or processes that could seriously harm or even eradicate human civilization on a global scale. These risks include nuclear war, pandemics, climate change, artificial intelligence (AI) safety, and biotechnology.
- Environmental Degradation: Humanity has been impacting the environment on a massive scale, leading to phenomena such as climate change, deforestation, pollution, and loss of biodiversity. These factors could have devastating effects on human life and the planet's ecosystems, leading to food and water shortages, migration, and other social and economic impacts.
- Nuclear War: The possibility of a nuclear war between major powers remains a significant concern. The destructive power of modern nuclear weapons could lead to catastrophic damage and loss of life, with long-lasting impacts on the environment and society.
- Emerging Technologies: Advancements in technologies such as AI, biotechnology, and nanotechnology could create unforeseen risks that could threaten human life and civilization. These risks could include the development of new weapons or the loss of control over intelligent systems, leading to unintended consequences.
- Unforeseen Natural Disasters: Natural disasters such as asteroid impacts, super-volcanic eruptions, and severe solar storms could cause significant damage or even threaten the existence of human civilization.
In summary, many potential threats could destroy humanity, including global catastrophic risks, environmental degradation, nuclear war, emerging technologies, and unforeseen natural disasters.
It mentions AI TWICE, and now I’m worried I’ve given it ideas…. Also, why do all AI’s start their final paragraph with ‘In summary’ or ‘In conclusion’, like they were twelve-year olds writing about crop rotation?
I then asked how gender equality could help mitigate these risks. Most of its response was what you’d expect, lots of stuff about women being in favour of environmental protection, and being less bellicose etc. But one point jumped out: “There is some evidence to support the idea that mixed gender teams make better decisions than single gender teams”.
SOME evidence to support?
Abandoning the AI for the moment I did some good ol’ fashioned research (or ‘Googling’ as we used to call it).
It turns out that the belief that mixed gender teams create better outcomes is widely, but not universally held. Those that argue that mixed gender teams have lower cohesion, and therefore lower success, have valid evidence and studies to point to in support of that view. So what is the truth? Enter Professor Letian Zhang:
In a study of 1,069 leading firms across 35 countries and 24 industries, we found that gender diversity relates to more productive companies, as measured by market value and revenue, only in contexts where gender diversity is viewed as “normatively” accepted. By normative acceptance, we mean a widespread cultural belief that gender diversity is important.
In other words, beliefs about gender diversity create a self-fulfilling cycle. Countries and industries that view gender diversity as important capture benefits from it. Those that don’t, don’t.
Now I had some interesting source material I went to Bing, asked it to reference Zhang’s work and build out some practical suggestions for taking advantage of it:
Based on Professor Zhang’s work on gender equality, one possible way for global companies to develop successful teams is:
- To promote gender diversity in their leadership and workforce, especially in sectors that require innovation and creativity
- To adapt their diversity strategies to different cultural and institutional contexts, as gender diversity may have different effects depending on the norms and values of each country and industry
- To leverage disruptive events as opportunities to challenge existing inequalities and foster a more inclusive and collaborative work environment
The first one is pretty vanilla, but the next two points sound interesting. Firstly here’s a bit more on cultural context:
One possible way for an organisation to adapt its diversity strategies to different cultural and national contexts is:
- To research and understand the cultural patterns, meanings, and values of its clients, partners, and employees in different regions
- To modify its policies, procedures, and interventions to be compatible with the local norms and expectations
Secondly, here’s more on disruptive events:
According to Professor Zhang’s paper, disruptive events are “unexpected shocks that disrupt an organization’s routine operations”. They can be external, such as natural disasters, economic crises, or social movements; or internal, such as leadership changes, scandals, or reorganizations.
Disruptive events can challenge inequality by creating three conditions:
- A sense of urgency that motivates people to act quickly and decisively
- A disruption of existing power structures that creates openings for new entrants and challengers.
- A shift in social norms that makes people more receptive to change and innovation
Some examples of disruptive events that have reduced inequality in organizations are:
- The 2008 financial crisis, which increased female representation in top management positions in US banks.
- The #MeToo movement, which increased female representation on corporate boards in US firms.
- The COVID-19 pandemic, which increased remote work opportunities and flexibility for women and minorities.
I pumped all this back in to the AI and asked for a summary sentence, it’s pretty good TBF (though, obviously it starts with ‘In conclusion…’):
In conclusion, gender equality is not just a moral imperative; it is essential for creating a sustainable and prosperous future for all.
But here’s the thing...
Professor Zhang’s work was carried out, documented, reported and widely discussed in 2020 and yet chatGPT (“My training data consists of a large corpus of text up until 2021”.) didn’t know who he was and was not aware of the significance of his work. How did this decision get made? Who decided that knowing the plot of Game of Thrones was more important than accurate, socially significant, information?
The risk is that, as more and more of the internet is created by AIs and then fed back into those AIs, we reinforce their ignorance and prejudices. We need to create a representative process for data inclusion. A process that allows diverse groups to put forward their stories, their information, and the data that influences decisions made about them. If we don’t take responsibility for what we teach the teacher, then equality will become nothing but a hashtag in a tech-bro echo chamber.