Fyma boss hits out at racial profiling retailers

Karen K. Burns, Co-founder and CEO of Estonian computer vision and artificial intelligence startup, Fyma, has called on retailers to stop asking shoppers for racial profiling data.

“Calling it ‘heritage data’ doesn’t make it OK, it’s still racial profiling. In addition to being illegal (as well as unethical), it’s technically extremely difficult to do,” Burns tweeted yesterday.

Shopping behaviour, she stated, is affected most by:

1. Gender: women shop more in bricks and mortar stores than men

2. Age: Millennials and high earners shop more online, older generations in physical stores

3. Income level - McKinsey divides shoppers into five segments that all shop differently

4. Weather

5. Other non-race related.

Retailers, can you please stop asking us for racial profiling data? Calling it ‘heritage data’ doesn’t make it ok, it’s still racial profiling. In addition to being illegal (as well as unethical), it’s technically extremely difficult to do. Shopping behaviour is affected most by:
— Karen K Burns (@Karenkangro) August 26, 2021

“On detecting race or ‘heritage’ - this can only happen in very controlled light conditions and needs to utilise face detection to work,” she added. “Meaning that in a regular shop environment it’s not possible. It’s also illegal in the EU and UK without the individual’s consent #computervision.”

Burns concluded: “Another issue is taxonomy - a standard taxonomy for human races was never established.”

“Computer vision generally works with seven of them, with datasets prone to various bias’ (capture, selection, category and negative bias). And again - not necessary for in-store analytics #AI.”

On a mission

With 770 million CCTV cameras installed around the world today and set to exceed one billion this year, Fyma’s mission is to help retail sector clients (retailers, shopping centre owners, real estate investment managers etc) unlock value hidden within existing video footage using AI and computer vision.

“We do this by transforming any new or legacy outdoor and CCTV camera into smart sensors that capture real-time data and extract insights,” Burns told RTIH in a recent interview.

“Our AI helps clients better understand the activity and behaviour of people/objects in a built environment to increase efficiencies and revenue.”

She added: “We offer a 100% plug and play SaaS solution that integrates with virtually any IP camera setup meaning clients don't have to invest in additional infrastructure.”

“Our platform is also self-learning removing the burden of manually inputting thousands of business rules and significantly reducing onboarding times.”

Fyma is funded by early stage VCs including Change Ventures, 7 Percent and Tiny VC. It is in the process of developing the business in the UK and other key markets in Europe.