With AI transforming our workplaces and the way we do business, the opportunities are huge – but so is what’s at stake. Experts like Tracey Spicer are urging more of us to be aware of how human bias can be unintentionally baked into the development of machine learning and, consequently, the AI technologies we may be rolling out into our businesses.
Tracey and Tarla led an insightful discussion around some of the critical topics Tracey covers in her latest book, Man-Made: How the bias of the past is being built into the future. The impetus for writing Man-Made came from a jarring conversation she had with her then 11-year-old son. After watching cartoons one morning, he asked for a robot slave – and in mimicking what he saw on TV, he prescribed his robot as female. What struck Tracey was this reinforced idea of women and girls being servile in the home, which we often see in games and on TV, and in some ways is being further embedded into the technologies we use in our everyday lives.
Whether conscious or not, humans all carry some kind of bias – even those responsible for programming and teaching AI. And with historical data sets informing algorithms, it’s important to acknowledge that historically women and many minorities have been disadvantaged. When asked about the implications of using historical data to inform AI, Tracey said, “Algorithms are opinions written in code." So, to prevent past discrimination from spilling into our future, businesses need to stay on the front foot by implementing things like robust auditing systems, safety by design principles and “human in the loop” systems, where people oversee the work of machines.
"Algorithms are opinions written in code."
For individuals, Tracey’s message is that we don’t need to be at the helm of tech developments to help mitigate these issues – she believes that end-users and businesses alike should choose to take some of the power away from the worst offenders. “Assess which tech companies you want to support, and investigate which companies have diversity and inclusion embedded in their workplaces.”
Despite a lot of analysis on the gaps in AI evolution, there are also opportunities for companies to strengthen their diversity and inclusion practices. Tracey believes the more we can work remotely may help us achieve greater diversity in leadership roles – particularly if organisations provide increased flexibility in their workplace practices to reflect the needs of their employees.
“Assess which tech companies you want to support, and investigate which companies have diversity and inclusion embedded in their workplaces.”
With many companies now using hiring algorithms to streamline recruitment processes, it’s imperative that these algorithms are implemented responsibly to avoid perpetuating bias and discrimination. We’ve seen, in recent times, examples of hiring tools discriminating against women due to historical data where this was prevalent – bots were throwing CVs of women in the bin because men were historically more successful in certain roles. “We thought it would prevent bias, when actually we saw the opposite,” says Tracey. Promisingly though, smaller tech companies are jumping in to correct biases in HR AI, with things like tools to help with incorporating inclusive words in job ads for businesses using hiring software.
No matter how we use AI, it's worth fine-tuning how we write prompts for better outcomes – and workplaces should consider investing in training to help their teams to get the most from the technology available. “The financial input is far outweighed by the benefits to be gained,” Tracey suggests, particularly when looking at AI’s capacity to take on the tedious elements of work. Taking away tasks like summarising long chunks of text or analysing large datasets can free up teams to support clients or customers, and give us time to be more creative.
Tracey expects that in the coming years we’ll see entirely new roles created due to the fast pace evolution of AI. We’ll also see existing industries change – some for the better, with huge potential health diagnostics and in caring roles, where robots can assist with aiding the elderly and people living with disability.
For business leaders, Tracey says, “You have more power than you realise. The dystopian narrative exists, but we have power even as one person.” Rather than looking at worse-case outcomes, incremental but solid changes towards a future of equality and inclusivity is possible.
What’s arguably the most critical factor that will determine the impact of AI on our businesses and lives more broadly, is the number of women, people of colour and people with diverse backgrounds in the field. Tarla and Tracey agree that greater scrutiny and safeguards around representation are vital if we want to make positive change. For business leaders, Tracey says, “You have more power than you realise. The dystopian narrative exists, but we have power even as one person.” Rather than looking at worse-case outcomes, incremental but solid changes towards a future of equality and inclusivity is possible – we just need to learn to master the tech, instead of letting it master us.
Tracey’s latest book Man-Made is available now online and in most book stores.
Sign up to the Women in Focus newsletter for more content like this.