Catherine Lückhoff, AI and cloud innovation entrepreneur

Catherine Lückhoff started her first venture, MANGO-OMC, while completing a Btech in Strategic Communication at the Cape Peninsula University of Technology, and her second venture, HQ Africa, in 2013. Drawing from her combined experience and her love for music, Catherine founded NicheStreem in 2015, an Amazon Web Services hosted end-to-end platform custom built to scale and support multiple white label streams.


También es la fundadora de 20fifty, una compañía de innovación en la nube e inteligencia artificial. Ha formado parte de la junta directiva de Silicon Cape Initiative y da charlas sobre diferentes temas: desde cómo el contenido propicia la adopción de la tecnología a los servicios africanos de música en streaming, pasando por la innovación en África. Su experiencia le otorga un punto de vista privilegiado en relación a la necesidad de estrategias de negocio descentralizadas.

Can you give us an overview of your work?

20fifty’s mission is to harness human ingenuity and intelligent technology to solve complex challenges, define new markets and unlock innovative and sustainable revenue streams both now and in the future. We are fortunate to work with forward thinking retailers, financial institutions, mining companies and IoT businesses, among others, for whom we architect rapid prototype and scale cloud innovation and AI driven solutions that work.

You've said that algorithms are better at recognising people than humans. How is that so?

The National Institute for Standards and Technology (NIST) shared a study of facial recognition technologies in 2018 that concluded that even older technologies could outperform human facial recognition capabilities. (See additional reference to this study on the AWS Rekognition facts page here) However, a more recent article published on June 4th, 2019 states the following:

“The best machine performed in the range of the best-performing humans, who were professional facial examiners. However, optimal face identification was achieved only when humans and machines collaborated.”

In my opinion FRT (facial recognition technology) is most likely better at recognising faces than the average (wo)man on the street, however when compared to the accuracy rates of trained facial recognition experts, (wo)man and machine are comparable. It’s a great use case for a human-in-the-loop approach of which I am a major advocate. Algorithms may be able to recognise an infinite number of faces without ever tiring and continue to improve over time, however having a human expert on hand to verify identifications when they matter most, is essential if we hope to preserve and protect civil liberties and privacy.

Is that so even with ageing faces?

With regards to the impact of aging on accuracy, I have no doubt that this does have an effect, however I can only assume that the impact on accuracy will depend on a number of factors not the least of which is the time that has elapsed between identifications.

Some people see face recognition more like a threat and a means of control than a tool designed to improve people's lives. How will face recognition make our lives easier and better?

To quote Elizabeth Barrett Browning in Sonnet 43, “Let me count the ways.” For starters, those already using Face Unlock on their mobile devices will agree that convenience is a major factor. Think of photo tagging on social media platforms like facebook, or the ease with which you can now board an airplane when facial recognition technology is in use.

More and more FRT will be used to facilitate access, to validate identity (for example at ATMs), to diagnose diseases, assist the blind or even help find missing children.

Yes there is risk of the technology being used as a tool to control as is evident by the application thereof in regions such as China and Hong Kong, however GDPR and the Facial Recognition Technology Warrant Act seeks to regulate the use of the technology by, for example, law enforcement and governmental bodies so as to guard against control. New technology should not be banned or condemned because of its potential misuse. Instead, there should be open, honest, and earnest dialogue among all parties involved to ensure that the technology is applied appropriately and is continuously enhanced.

Experts are warning about face recognition technology being gender and skin biased. You have said that there should be enough diversity in facial recognition as to prevent this from happening, but how is that possible if there isn't enough diversity in the use and availability of technology in general?

In my opinion it is less about the use and availability, and more about the creation and training of the technology. As with any algorithm, it is only as good as the data it is being trained on, so it is imperative that the creators of the tools ensure that they both plan and audit for representation in their data. For those using and actively engaging with the tech, we are all responsible for applying context limitations to systems we develop to mitigate harm. And it goes without saying that everyone should be schooling themselves on best practices.

A bill introduced last month in the US would ban users of commercial FR technology from collecting and sharing data for identifying or tracking consumers without their consent. Is this a widespread use of this technology?

I cannot speak to the implications of the bill as I do not have a legal background, however,

I do believe that the technology for identifying and tracking consumers is more widespread than people realise.

The use of online data ( e.g. search history, page visits, purchasing behaviour, clicks and more) as well as biometric data such as age, gender and mood is all being applied to target consumers with “smart advertising” both online and off. I confess it concerns me as transparency is seriously lacking and consumers may not even be aware that their choices are being influenced.

Are you a social media user? If not, why?

Not particularly, no. In fact in recent years I have deleted most of my social media accounts with the exception of Twitter.

The older I get the more I value my privacy and I simply don’t trust the tech giants with my data.

I also don’t need to live vicariously through the timelines of other people and I believe the trend in social media that drives people to seek validation through “likes” and comments is a slippery and lonely slope. I am very fortunate to have a close group of friends and family to whom I turn for advice, love, support and comfort and I prefer spending my time with them in person rather than engaging online.

San Francisco, long at the heart of the tech revolution, has banned the use of FR. Is there a lack of legislation and guarantees as to protect civil and human rights from this technology?

At a global level I do believe so yes, however from country to country and state to state this topic is being addressed to a greater or lesser degree. Culture, safety concerns vs. privacy concerns, the political landscape, socio-economic factors, connectivity, education and more all play a role in how, and how soon, countries are adopting their own legislation and the protection this offers their citizens. How and whether the use of this technology can or will be legislated on the international stage remains to be seen.

You also talked about people already working on technology to deceive face recognition technology such as patterns in clothing. What about plastic surgery, will we see the raise of plastic surgeons specialised in face recognition deceiving surgery?

I have read a number of articles recently that address this topic and am pleased to see the likes of the Journal of the American Society of Plastic Surgeons add their voice, however changing one’s appearance to evade the Law is certainly not a new subject.

The number of data points used to create a unique ID are numerous, so I can only assume the transformation would have to be significant.

For now wearing a mask is probably a less risky and costly endeavour.