women need to be heard at the AI table

Katya Moskalenko is the product marketing manager of Measure Protocol in London. She warns that artificial intelligence could have a negative effect on the gender gap already present in the tech sector. It opens up a whole new world of possibilities but also carries with it – almost by default – the risk of bias and exclusion.

She told us that to avoid these traps “a whole series of decisions and strategies will be required, such as ensuring diversity in datasets, diverse teams, and ethical considerations for any AI-powered solution.”

There is a long road ahead. Moskalenko quotes World Economic Forum data from the year 2021, which suggests that only 26% (or fewer) of AI and data positions are held in by women. There’s no reason to believe that this statistic has improved dramatically.

It’s going be a long trip

Katya acknowledges that there is no quick fix. She said, “We cannot solve everything all at once.” It’s going be a long road to ensure AI serves all individuals. We can make sure that new technologies like AI are fair and impartial by having a balanced representation.

The time is now to begin. She said that it sometimes appears as if every company, large and small, has an AI team working on new solutions.

She continued, “So let’s first diversify our teams by adding more women, more people of colour, and more minorities to ensure that we have a diverse, inclusive perspective.” “Also let’s make sure we have diverse and inclusive datasets to feed into systems.

The fear that AI models, particularly those trained online, may develop biases , is not a new concern. There’s no simple solution to this problem, though Moskalenko admits that the major players are trying.

She said, “I am quite impressed with OpenAI’s guidelines. They give good instructions on being cautious and responsible.” “It’s also important to share the best practices and some of the vulnerabilities and weaknesses we find. It is important to mitigate unintentional injury.

She calls for cooperation. We need to admit that there are problems and work together to resolve them. “With all of the economic and commercial rivalry going on, we should think about society as a unit, humanity as an entire. We will live in the society we create and the humanity that we develop. “Effective cooperation is key.”

Get those datasets correct

Theresa Kushner, MarTech contributor and data evangelist, is also heavily invested in inclusion and diversity when it comes women in AI. We asked Theresa Kushner about the link between diverse datasets and diversity in teams.

She emphasized that “ensuring gender equality begins long before you design AI algorithms.” It all starts with the data that you collect. It’s interesting to note that most companies do not always think that gender capture is a necessity. They can’t say if the data they have is biased. I’ve worked for companies that have inferred gender based on information such as names, college affiliations, and extracurricular activities. It’s not the best way to determine gender but it is what they have.”

She also agrees with the need for a diverse team to evaluate AI algorithms. As data becomes more of a product than a service, AI teams need to think like product developers and consider their users and markets. An engineer once told me about the team who created a Fitbit like product. The designers of this tool, all men, forgot to include the only thing every woman is interested in tracking – her period. We would have the same situation if we didn’t have diverse groups producing data products.

Dig deep: Why AI is important in marketing

Retaining employees is important, but so is hiring.

Hiring for diversity alone is not sufficient to diversify the teams in this field. There should be initiatives that promote not only hiring, but also retention. Because unfortunately, it is not enough to hire a female. It’s important to ensure she’s up-skilled or reskilled.

One thing is for sure: if we put more emphasis on retraining male workers to be able to work in generative AI, and AI and data in general, this will only worsen the situation.

Kushner agreed with the statement and elaborated on it. “Ofcourse, you should both train in AI, but we shouldn’t be concerned about the tools. Women in IT should have easy access to tools. But we should also ensure that women participate in the governance of these tools and AI algorithms. Diversity isn’t just about gender or ethnicity. It also includes diversity of approach and thought. “Including women gives you this kind of diversity.”

Women in tech: a global movement

Moskalenko’s involvement in Women in Tech, and not just AI, has allowed her to be involved in these issues. She explains why.

She told us, “I have had a very interesting journey from traditional marketing fields to fast-paced technology startups.” (Measure Protocol provides software that tracks competitive intelligence and consumer behaviour). I felt it important to share my experiences and help other women explore this vibrant and interesting industry.

You can find out more about Women in Tech’s global mission here.

The post Women must be heard at the AI table first appeared on MarTech.