Is tech bias entrenching existing prejudices?

An in-depth look at whether there is a tech bias in AI and if it is entrenching existing prejudices in society with possible solutions.

Raabia Fazil | 10th Jan 20
game pieces arranged together with a darker one separate from the group symbolising if tech bias is entrenching existing prejudices?

AI is cool. The term evokes slick machines automating processes and making our lives easier and smoother. However, is AI empowering everybody to benefit or is tech bias entrenching existing prejudice?

An extensive branch of computing, AI seeks to build smart machines to complete tasks requiring human intelligence. Wendy Hall, a professor of computer science at Southampton university, has contributed to a report on artificial intelligence for the UK government. ‘With AI’, she says ‘we talk about bias in, bias out.’ As artificial intelligence is deployed in recruitment, biased input can generate biased output, adversely affecting women, disabled people and ethnic minorities.

Michael Sippitt, a director of Forbury People, a HR consultancy, says ‘tech races ahead of people working out how to use it’ adding that AI systems are built on historic data so are likely to select people similar to existing teammates instead of addressing under-representation. Kriti Sharma, vice-president of artificial intelligence at Sage, says ‘A lot of the CVs and historic profiles will be of one kind of candidate…If you were hiring a chief technology officer for a company and the algorithm was learning from historical data sets then what would you expect?’


> Read more: Research shows that 48% of charities want to use data to improve service delivery


The current landscape and a chance for change

Research published by the AI Now Institute, a New York University research center, has found that AI, largely developed by white and male techies, is highly likely to perpetuate gender and racial biases, along with accompanying power differentials. The report found include image recognition software making offensive classifications of minorities and chatbots adopting hate speech. Kate Crawford, an author on the report, said ‘The industry has to acknowledge the gravity of the situation…The use of AI systems for the classification, detection, and prediction of race and gender is in urgent need of re-evaluation.’ The report found that over 80% of AI professors are men whereas only 15% of AI researchers at Facebook and 10% of AI researchers at Google are women.

Tess Posners, CEO of AI4ALL, an organisation seeking to raise diversity in AI says currently there is a chance for techies to take a shot at the problem as AI grows and expands, she said ‘right now we are in an exciting moment where we can make a difference before we see how much more complicated it can get later.’


> Read more: Solving humanity’s biggest problems with Artificial Intelligence: a quick guide


How can the situation be tackled?

Unconscious bias is difficult to recognise and therefore difficult to alter. A lot of organisations can have a corporate bias based gender or culture without being aware of it. But the ones championing inclusion stand the strongest chance of success in the future. As companies carefully weave artificial intelligence into their development, how they construct their teams should be equally considered. It is notable that greater diversity brings greater profitability.

Diverse teams can magnify the inclusive capacity of AI to recognise a wide range of people. AI itself can be expanded and developed based on how humans build it. There are a few ways to do so:
Spreading awareness of AI and understanding the diverse requirements of users then including them into the design of AI can help to expand its remit. Attracting women to work in technology and use AI can help too. Ensuring that inclusion is embraced on every level of employment structures across industries including IT services and finance can help to expand the limits of perception.


> Read more: Making success out of failure – learnings from charity digital campaigns


How is the charity sector addressing bias and prejudice?

We reported that the charity sector is shaking up its approach to inclusion. We reported on the #NonGraduatesWelcome campaign urging charities to hire people on merit, experience and skills instead of selecting only graduates. It follows the #CharitySoWhite Twitter campaign, launched to highlight racism and promote inclusion. The government has introduced a law for all organisations with over 250 staff to submit their annual gender pay gap data to address the differences in salaries and proportions of men and women in different pay grades.

The significance of inclusion in the charity sector has been highlighted by the massive response to the #CharitySoWhite Twitter campaign. Our podcast episode titled ‘Unprepared Vs. Prepared Organisations‘ explores the ways in which charities can embrace inclusion. This episode delves into the issues raised by homogeneity and the many advantages of embracing inclusion and how it can champion progress in the sector.