Gender Diversity in the Era of AI

Navigating gender bias in the corporate world – the impact of AI on gender diversity targets

A follow up from the WeQual Think Tank, The AI Challenge, with a panel of our WeQual Alumni industry experts.

Artificial intelligence (AI) in business isn’t just about smarter tools and increased productivity; it has the potential to reshape how businesses run, including how we address gender dynamics.  

At a recent WeQual Think Tank event: The AI Challenge, (watch the recording of the event below) industry leaders Christina Montgomery, Chief Privacy and Trust Officer at IBM, Vrinda Menon, Global Head of Managed Accounts & Client Services Technology at JP Morgan, and Hanna Helin, Global Head of Technology Innovation at LSEG, discussed AI’s profound impact on women in business, and sparked some enlightening conversations. 

In an organization of just 100 people, if you introduced an AI solution for each individual, imagine how much more work they could do in the same amount of time.

The Dual Face of AI in Business

There’s no doubt AI is a game-changer – automating mundane tasks, optimizing complex processes, and unlocking new insights through data analytics. “Right now, any person in the world who wants to experiment with AI, who wants to learn, can go in and do something with tools like ChatGpt,” says Vrinda. “And every piece of work that anybody does could eventually be made better with AI. Therefore, I think every single job is going to change dramatically.”

In an organization of just 100 people, if you introduced an AI solution for each of those individuals, imagine how much more work they could get done in the same amount of time. “I think it’s a huge opportunity,” says Hanna. “Even from a societal perspective, thinking about these comments we’re hearing from the market, like talking about a four-day working week, for example – I think it opens up huge opportunities in the market and really adapts the whole workforce going forward.”

New jobs are being created, too. Christina points out: “My job has changed dramatically with the introduction of AI and a lot of the folks on my team who work in AI governance, these were not roles that existed until very, very recently.”

Here’s the thing: AI’s brilliance doesn’t entirely blot out its darker potential consequences – gender bias being one of them.

AI will soon be woven into the fabric of your organization, and that means those in the top decision-making roles need to be well-versed in the new technology.

AI and Gender Bias

Consider the AI-driven recruitment tools designed to streamline hiring processes. While they can efficiently pare through thousands of applications in a short time, if not calibrated correctly, they risk favoring the same demographic profiles for certain roles that humans may have. This was observed with companies like HireVue, which had to recalibrate their algorithms after realizing they might unintentionally favor male candidates.

These biases don’t exist just for candidates; AI can inadvertently stereotype roles based on historic norms – for instance, virtual assistants and chatbots are often designed with female personas, which perpetuates traditional gender stereotypes of women as helpers or caregivers.

Meanwhile, in industries like administration and retail – which typically have a higher proportion of women in their workforce – AI automation is causing job losses. This further widens the overall gender gap.

So, how can we bias-proof our AI systems so we don’t find ourselves slipping back a few decades on the gender needle?

Future AI models are likely to be more efficient but also more interconnected.

Practical Solutions for Gender Diversity in AI

AI will soon be woven into the fabric of your organization, and that means those in the top decision-making roles need to be well-versed in the new technology. They will need to set a course and lead the charge for the organization inclusively and responsibly.

Establish Clear AI Governance Policies

Begin by setting up an AI governance framework – what does responsible AI use mean for your organization? How will you manage your data and how will you ensure accountability? This framework should outline the specific responsibilities of all AI stakeholders, from IT teams to department heads, ensuring that everyone understands their role in maintaining AI integrity.

Christina, who directs IBM’s privacy and AI governance programs, says: “We started with establishing a set of principles and values around where we were going to use AI and the guardrails we were going to put in place when we were using it, and we built them into our practices across our company through mechanisms like an AI ethics board and the oversight for that board by senior management. Your governance program needs to be evolving as well, all the time, and our products now are all geared towards enabling our clients to build trustworthy AI.”

Implement Comprehensive Bias Monitoring Systems

It’s crucial to continuously monitor and test AI systems for bias. Organizations should develop mechanisms to regularly assess AI decisions against fairness criteria established by your governance policies. This can involve routine audits by internal or external experts. Tools like AI fairness metrics and dashboards can provide real-time insights into how AI applications perform, helping you identify and mitigate biases swiftly. Hanna says: “We have formed a working group, along with our risk, legal, business, and technology teams to establish responsible AI practices within our organization.”

Invest in Diverse Data Sets

Programmed well, AI can work in gender equality’s favor. It can unearth hidden gems within organizations by analyzing performance data to help identify potential leaders, especially women, who might be overlooked.

Bias in AI often stems from incomplete or non-representative data. To counteract this, variety is a key. Invest in diverse data sets that reflect the broad spectrum of your customer base and operational environments. Encourage your data teams to source data from a variety of demographics and geographic locations. Moreover, involve subject matter experts during the data curation phase to ensure the data’s relevance and inclusivity.

“We really need to make sure we bake in responsible practices for the organization,” says Hanna. “We were leveraging many of our existing practices, things like model risk management, and basically baking in those AI-related questions, there-and-then, at the same time.”

Foster an AI-Literate Workforce

Promote AI literacy by providing training programs that educate employees about the basics of AI, the importance of unbiased algorithms, and their role in supporting ethical AI usage. This should include training on recognizing and reporting potential biases they might encounter in AI outputs.

For example, IBM has developed a set of courses called IBM Skills Build that are freely available in areas like AI, cyber and data analysis and cloud computing. With over 1000 courses in 20 languages, they are particularly focusing on underrepresented communities. Christina says, “The last thing that we want is for the AI evolution essentially to leave behind those that need it the most.” An informed workforce can act as a first line of defence against biased AI decisions.

Encourage Transparency and Stakeholder Engagement

Maintain transparency about how AI technologies are deployed within your organization. This includes clear communication about the workings of AI systems, the origins of the data they use, and the measures in place to ensure fairness. Engage diverse and inclusive groups of stakeholders – such as employees, customers, and industry partners – in discussions about AI applications to gather a wide range of perspectives and feedback. This not only builds trust but also enhances your AI systems through collaborative input.

How can women distinguish themselves?

If you’re a woman leader looking to stay abreast of AI technology, the key is to keep up to date. AI is a leveler, and women leaders have the opportunity to really think outside the box.

“Every single day, I think we need to think about how every piece of work could be done differently with AI,” says Vrinda. “We need to think big and transformational. There are new products, there are new research papers and new ideas and applications of AI coming out frequently and, so iterate small, dream big.” 

“The positive point with this trend is that there’s so much information available,” says Hanna. “Follow thought leaders on LinkedIn, and take advantage of free courses from cloud providers, there’s a flow of information out there.”

“You don’t have to be a technologist or a data scientist in order to be a leader in AI right now, and to understand it and to help lead,” adds Christina. “Women have been leading the charge on responsible AI innovation and research in governance around that space for years. So, when we talk about developments and advancements in AI, those that are the technologists really need to make sure we have that aperture of what we mean by responsible AI development. It’s more than just the technology of the largest model, it’s how do you deploy that largest model with trust, how do you test things for bias?”

So, what might the future of AI hold for big businesses?

Multimodal Large Language Models

According to Hanna, the future could involve the integration of text, images, and voice into powerful, multimodal large language models. These models will aim to process and understand multiple forms of data in a way that mimics human sensory and cognitive capabilities. “Really using text, images, voice and that all coming together in these powerful models,” she says. This convergence allows for more natural and intuitive interactions between humans and machines, facilitating a broader range of applications from automated content generation to sophisticated personal assistants.

AI and Sustainability

A significant emerging trend is the cohesion of sustainability within AI development. Hanna notes that there is a growing conversation about moving from large, general-purpose models to smaller, more specialized models that require less computational power, thereby reducing their environmental impact. She states, “Do you actually need to train the full model or just focus on specific piece of that model… it will also lead to the point that you don’t need as much computing power for your task.” This shift not only optimizes resources but also addresses the critical need for sustainable technology practices.

Interconnected AI Systems

Future AI models are likely to be more efficient but also more interconnected. Hanna suggests improvements in how different models, such as language and vision models, can work together seamlessly: “Various different models connected in a much more efficient way.” This interoperability could enhance AI’s ability to handle complex, multi-faceted tasks by leveraging the strengths of different models in concert.

Artificial General Intelligence (AGI)

AGI is undoubtedly the pinnacle of AI research aiming to create machines capable of general human-like intelligence. Hanna says, “Artificial general intelligence – we are not there yet, it will take time before we will be there.” The development of AGI involves addressing numerous ethical, technical, and societal considerations before such systems can perform any human task independently. But just imagine a future where technology and humans work in symbiosis. Both scary and exciting, right?

AI Agents

Vrinda introduces the emerging concept of AI agents that could dramatically change the landscape of work and automation. These agents would not just perform isolated tasks but could manage complex sequences of actions across different domains. She explains, “The notion of agents, where you can define a piece of work and with reasoning and action, that agent is able to break the work into tasks and orchestrate those tasks in a sequential or a parallel way to completion.” While early days, as these technologies and capabilities mature over time, this idea suggests the future potential for more integrated AI systems capable of managing entire workflows and processes.”

In Summary ...

AI is a transformative force that can redefine gender equality in the workplace. However, leveraging AI in a way that truly benefits all requires careful planning, continuous oversight, and a commitment to fairness. The insights shared by leaders during this WeQual Think Tank paint a clear picture: while AI presents challenges, with the right strategies in place, it also offers substantial opportunities for a more inclusive and equitable corporate world.

 

Watch the event video:

Thanks to our 3 Guest Contributors:

Christina Montgomery

Chief Privacy and Trust Officer at IBM

Christina is also a WeQual Alumni and Winner of the WeQual Awards, The Americas 2021 in the Commercial Innovation category.

Vrinda Menon

Global Head of Managed Accounts & Client Services Technology at JP Morgan

Vrinda is also a WeQual Alumni and Winner of the WeQual Awards, The Americas 2023 in the Digital & Technology category.

Hanna Helin

Global Head of Technology Innovation

Hanna is also a WeQual Alumni and Winner of the WeQual Awards, EMEA 2023 in the Technology category.