Nozha Boujemaa of Median Technologies: ‘We need to build a secure AI system’
The science and innovation director at medical imaging group Median Technologies, Nozha Boujemaa has become a leading figure in artificial intelligence (AI), advising organisations including the European Commission. In an interview, she talks about the technology's potential for Africa.
Today, can all companies use artificial intelligence (AI), regardless of their size or sector of activity?
Nozha Boujemaa: Yes, in fact the difficulty does not come from AI itself but from the data. People have not yet understood how important their structuring is. A company that wants to use AI must be able to exploit data even if they are multi-source, so they must be structured. We always talk about algorithms, but the algorithm is only the engine. Data is the fuel. Then there is the issue of skills. To deploy an AI system, a software engineer is not enough, you really need to have a fairly advanced knowledge of data science.
To illustrate this, how do you use AI in your company, Median Technologies?
Our sector is personalised medicine, and more specifically medical imaging. We extract indicators that allow professionals to make predictions. We work with radiologists and help them in their decision-making. In concrete terms, when they are not completely sure how to interpret their data, they have a search engine that will study a considerable number of similar cases and give them statistical information – in such and such a case, this was the diagnosis – to reach a decision. We also work with pharmaceutical companies. We help them save time and money through therapeutic innovation and by allowing them to seek prognostic information on “responder patients”. This is very important in immuno-oncology, where we know, for example, that some treatments are only effective in 20% of patients.
AI is neither a magic wand nor something intrusive or dangerous in itself.
Is it an additional tool, as in an improvement on the existing one, or is it really a revolution?
It’s not just another technology, it’s essential. Humans can’t process the amount of data they have on their own. We must therefore adapt, deploy this technology while keeping control over it, and the danger is to over-sell it. AI is neither a magic wand nor something intrusive or dangerous in itself. What is needed is to put in place safeguards to build what is called a “secure AI” system: one that is robust, that respects ethical values, does not degenerate into forms of unintentional discrimination or political manipulation, as we have already seen.
So specialists like you are asking questions about social acceptance and regulation for AI?
Of course. That is why, before joining Median, I founded the DataIA Institute, an interdisciplinary organisation that brings together technology specialists, sociologists, philosophers and lawyers. There are always debates between techno-sceptics and enthusiasts who do not want to regulate anything. For me, regulation is not the right answer, but we must ensure a certain transparency, be aware of responsibilities and be vigilant in managing any risks that may arise.
Will AI destroy jobs?
Some people say so, but I don’t believe it. Jobs will disappear, but new ones will be created. In this field, too, AI is transforming, and that is all the better. It will replace human intelligence for repetitive tasks and allow us to devote our minds to more important things.
We must provide the means to make the brightest people want to stay.
Have you seen any interesting initiatives emerge in Africa in these areas?
I have met some extremely bright people. There are already very good training courses on the continent and, thanks to the internet, students are not limited to the courses available on the spot, they are also trained online. The breeding ground and training are there; the demography is favourable. The only concern I have is that the talent will go abroad. We must provide the means to make the brightest people want to stay. As for the areas to be explored, I have already discussed them with some African colleagues: they need to collect data on the issues they face. Infrastructure, water, environment and energy.
Data is the key, I repeat, it’s a source of wealth. All this can open up very large markets, especially in infrastructure. But there’s a lot of work to be done.
This article was first published in Jeune Afrique