Women and AI: How startups can prevent gender bias and promote responsible use of the tech | e27

Despite progress, there continue to be challenges in ensuring a level playing field for people of all genders in the startup ecosystem.

“While evening out the playing field via representation numbers is a helpful indicator of progress, this isn’t the all-encompassing practice to promote gender equality. The core of closing the gender gap stems from fostering the right support system and environment to retain and nurture bright talent, while our leadership models the type of values we expect from each of our employees,” says Veronika Folkova, Senior Director, People Business Partners for Global Legal Organisation and APAC, Confluent.

Silicon Valley-based data streaming platform Confluent has implemented initiatives to ensure equal opportunities for employees of all genders. However, with the rising popularity of AI technology, as the technology becomes more widely used, most of us would start to ask questions: Is there any risk related to gender equality in this phenomenon?

How will the rising popularity of AI affect gender equality in the workplace, particularly in the tech industry? How can startups play a more active role in promoting the responsible use of technology? In this email interview with e27, Folkova discusses the issue.

The following is an edited excerpt of the interview.

Can you share with us how gender bias can show up in AI? What potential/existing harms can it cause?

Gender bias within AI is quite a complex topic in and of itself. As AI pulls information by ingesting large amounts of training data, analysing it for correlations and patterns, and using these patterns to make predictions for the future, its input and output impact people across different backgrounds, industries, and socioeconomic levels.

Also Read: Revolutionising retail: A blueprint for future success

The truth is that women in many countries have limited access to training and education, even more so in the field of R&D, which may result in unconscious biases. Fundamentally, most content and research on AI are typically written in a language that automatically assumes the reader’s perspective as a man.

This forms the foundational bias for words used and embedded into algorithms, which leads to misinformed algorithms and ultimately breeds gender discrimination and exclusion of qualified individuals.

One real-life example is the Generative AI art generation application Midjourney. When asked to create images of people in specialised professions, it showed both younger and older people, but the older people were always men, reinforcing the age-old bias against working women.

What are your proposed solutions?

To be honest, there isn’t a one-stop solution to gender bias in AI. From my capacity as Senior Director, People Business Partners for Global Legal Organisation and APAC, we are seeing trends where AI may play a significant role in the future of sourcing candidates and coordination of recruitment processes.

As such, it goes back to laying the groundwork to provide equal opportunities to the right people, doing our own thorough research to uplift the right talent, and improving the training and development of our own people to identify the limitations of AI and better use its strengths.

How can the government take a proactive role in this?

In the recent Budget 2024, Deputy Prime Minister Lawrence Wong announced that Singapore will invest S$1 billion over the next five years into AI computing capacity, talent and industry development.

This will support Singapore’s National AI Strategy 2.0, which was launched in December last year.
These investments are in the right direction, coupled with the growing awareness of AI’s shortcomings, such as biased data and hallucinations, i.e. AI model learning incorrect patterns that lead to incorrect predictions.

Also Read: The future of AI-driven tech solutions in enhancing business infrastructure

To combat this, I appreciate that the Singapore government has set up a Generative AI framework that outlines a “systematic and balanced approach” for Generative AI concerns while facilitating innovation. We need to keep working together, educating each other, and pooling resources to keep up with the advancements in technology and understand how it can be used for good for the wider population.

What role can tech startups play in ensuring responsible use of AI?

While they might be smaller in size, startups have the ability to move the needle in the industry and influence change in their area of expertise.

Some ways that they can ensure the responsible use of AI include:

– Diverse and inclusive teams for bias mitigation: Startups should prioritise building diverse and inclusive teams that bring together individuals with different backgrounds, perspectives, and expertise. This will ensure teams can create AI systems that account for diverse needs and backgrounds and identify and mitigate biases in AI outputs.

– Transparency and explainability: Startups can promote transparency and explainability in AI systems by clearly documenting how their algorithms work, including data sources, training processes, and decision-making criteria.

– Collaboration and knowledge sharing: Startups can contribute to the responsible use of AI by actively participating in industry collaborations and sharing best practices as they test and run new ways of doing things.

As a business, do you have any initiative to prevent the harmful use of AI?

As Confluent drives data streaming with real-time data, we ensure our customers can tap into continuously enriched trustworthy data streams to quickly scale and build real-time AI applications for maximum value.

Also Read: From peak scrolling to personalised communities: The Gen AI solution

Data streaming is quickly becoming the data backbone for modern AI. Our solutions can consolidate an organisation’s operational and analytical data from disparate sources to construct a unified source of real-time truth of all their data. This not only empowers them to excel in model building and training, driving unparalleled levels of sophistication and accuracy across various applications but also helps identify any signs of abuse as soon as possible.

—

Image Credit: Confluent

This article was first published on March 22, 2024

The post Women and AI: How startups can prevent gender bias and promote responsible use of the tech appeared first on e27.

Source link