Side face of AI robot by particle form.
Yuichiro Chino | Moment | Getty Images
The global pandemic has aggravated existing inequalities around the world, and triggered questions about whether technology can help level the playing field.
As a leading hub for cutting edge technology and innovation, Asia faces the same debate: Can technology work for all?
Developing economies are more vulnerable to the inequities of access that exist in technologies like artificial intelligence, according to Kay Firth-Butterfield, head of AI and machine learning, at the World Economic Forum (WEF).
“That’s partly because we don’t have sufficient data, because it’s simply not being created,” she told Rosanna Lockwood during a special episode of CNBC’s The Edge.
AI can really be used in every area, where we need to make progress as human beings. So whether that’s on climate change, or health care, or an education.
Kay Firth-Butterfield
head of AI and machine learning, World Economic Forum
“We need data to train the models and we need that data to not be prejudiced,” she said, adding that developers need to come from those same emerging economies.
“There’s also a problem that so many people don’t have internet and don’t have access to the tools of AI. So these developing economies are being prejudiced because they can’t access AI, and the benefits from it,” she added.
Debiasing A.I.
Despite its enormous potential, AI faces several challenges. It has been widely criticized for perpetuating inequalities due to its inherent biases.
For instance, Firth-Butterfield pointed out that most of the developers who write the programs are men — that means women are underrepresented.
“Many of the people who are training the algorithms tend to be men and tend not to be very diverse in their backgrounds. So that means that they bring to the training of the algorithm that lack of diversity,” she said, adding that often means the input data can contain the historical biases of those who created the data.
On top of that, AI’s facial recognition software can come across discriminatory as it does not recognize certain races.
“What we have seen is, for example, that it has caused very poor outcomes in facial recognition for Africans. We have seen that it has caused very poor outcomes when we use AI for loan development, or when we’re using AI to help judges decide on bail applications,” Firth-Butterfield noted.
“So we really need to work on curing that. And one of the ways you can do that is actually just to make sure that you have diverse teams around the the developers.”
Elevating women in technology
The pandemic has also exacerbated inequalities like gender and race-related problems across Asia.
Celine Le Cotonnec, chief data and innovation officer at the Bank of Singapore, pointed out that Covid-19 has revealed huge inequalities across different businesses — and that includes hampering the progress of women in society.
“Everybody had to work from home. The kids were not going to school, women being the primary caretaker of kids they were the first ones that actually had to take a step down in their career and in their activity,” Cotonnec said.
“I believe that it has hindered the progress that the society was actually making,” she added.
Bank of Singapore recently joined the SG Women in Tech initiative, a government-led effort that aims to inspire women to consider a future in the field of technology across Singapore.
“We have pledged about 30% of our new fresh grad hire would be female. A second initiative is about empowerment of the woman,” Cotonnec said.
“So we are running … a workshop on empowerment of women in technology — for them to be able to say: ‘It is feasible, I am about to actually not only work in technology, but tomorrow, take a leadership role in in our organization,'” she elaborated.
Government efforts
As part of its efforts to accelerate the benefits of AI and technology, the World Economic Forum launched the Centre for the Fourth Industrial Revolution in more than a dozen countries. The goal is to work with governments and businesses to help mitigate risks and test frameworks for governing technology.
I would say that governments do need to have some national AI strategies and they need to do that urgently.
Kay Firth-Butterfield
head of AI and machine learning, World Economic Forum
“AI can really be used in every area, where we need to make progress as human beings. So whether that’s on climate change, or health care, or an education. So really, we want to maximize these benefits,” said WEF’s Firth-Butterfield.
“I would say that governments do need to have some national AI strategies and they need to do that urgently. There are not many in the developed [and] developing economies,” she said. “Obviously, the tech companies have a role to play. Companies are not spending money in the way that they could or should be doing.”
She cited India and Singapore as examples of countries working on developing a national AI strategy. Such a blueprint allows businesses to know what the government is planning to do in this in this area, she said.
“Singapore has really been showing the way. They worked with us on their model, governance framework for AI, and how companies should deploy AI and think about the ethical challenges,” Firth-Butterfield noted.