India currently has no AI-specific laws. The personal data protection bill (2019) came closest. But that too was withdrawn by the government as the Joint Parliamentary Committee proposed a variety of changes and decided to call for a new law rather than amending the existing one.
Analytics India Magazine interacted with Mariagrazia Squicciarini, Chief of Executive Office and Director of AI at UNESCO. She suggested some aspects in drafting AI policies and reforms for India and shared some of the challenges, solutions, the role of government and more.
Highlighting one of the cases from a political point of view for India, she said: “Theoretically, if you provide your data, in a year you have the possibility to remove/delete it (data). But in practice it becomes extremely challenging. So the question for India and many other countries is how to actually implement some of the recipes.”
She said that India should consider having a harmonized policy that has the ability to go global in activities, rather than each country having different approaches that actually limit the possibility of compliance by companies, start-ups and customers using technology use, affect.
“I hope they take this into account in development now, and I’m sure they are,” Squicciarini added, saying they may reflect several other limitations in the final formulation.
At UNESCO, Squicciarini’s work includes strategizing and advancing the organization’s work related to social change and the reduction of inequalities – human rights, anti-racism and gender gaps – the ethics of science and technology, and particularly AI and neurotechnology.
Prior to that, she worked for over a decade as Senior Economist – Head of Unit at the OECD Directorate for Science Technology and Innovation. Her work revolved around digital transformation, artificial intelligence and others.
Squicciarini defines AI ethics as “putting technologists at the service of humans, not humans just to be used by technologists” and ensuring that humans are at the center based on human rights.
India vs World
Last year, UNESCO announced that all of its 193 member states had adopted the Global Framework Convention on the Ethics of Artificial Intelligence, which emphasizes the benefits of AI technology to humankind and raises fundamental ethical concerns. Recently, the US, which is not a member of UNESCO, also introduced the AI Bill of Rights, which aims to prevent the damage caused by the rise of AI systems.
Slowly but surely, many countries are now realizing the new reality created by artificial intelligence and other emerging technologies before it is too late to act. According to the AI Index report, the UK and US topped the list, following the pattern of the proportion of AI mentions in legislation that became legislation. Great Britain comes first with 939 entries, followed by Spain, Japan, the USA and Australia.
On the contrary, the number of mentions of AI in Indian legislative processes in legislative processes between 2016 and 2021 was 34. Out of 24 countries mentioned in this list, India ranks sixth.
(Source: Bloomberg Government)
Challenges in drafting an AI policy
Squicciarini said that today many companies implementing AI call themselves differently. Most companies today say they use AI and machine learning, but things like convolutional neural networks aren’t often mentioned. In your opinion, this remains a challenge as to how we can translate this into policy.
She said once that’s done, how do you put principles into practice? “Because the challenge for AI is that if you look at the recommendations, it touches so many parts of governments. We know governments often work in silos simply because it’s very complicated to get everyone on board,” Squicciarini said. She said: “Let’s face it, coordination isn’t easy. The coordination effort is high.”
Citing health services, she said the impact is being seen health to occupation to privacy. “Of course, what we’re already talking about with the example – I put in place three ministries,” she added, saying they need to regulate and legislate in the same direction. You must enforce this in a way that ensures consistency. “So those are the challenges that we see,” Squicciarini said.
What is the solution?
She said that there is a lot of fermentation and a lot of initiatives around the world. This is both good and bad news. “We shouldn’t reinvent the wheel; We should check what works and share best practices,” she added.
Squicciarini said that countries belonging to the same region should work together because they share some similarities. This can be used to test or experiment in one direction. Learning and sharing together helps reduce costs, save time and improve the efficiency of the whole process.
Citing Germany and Canada, alongside Namibia and other South African countries, Squicciarini said these countries have seen multiple efforts and are standing together to drive implementation forward. “So there’s a lot of good practice emerging,” she added.
Squicciarini’s version of AI politics is designed to find a global solution to a global challenge. “In the sense that AI has no limits, if you don’t set them, you can’t stop them,” she added, saying that there’s no point in having individual, wildly different types of regulations or strategies.
She also said they see the solution as intergenerational, as more and more young people are demanding the solutions and the older generations, who sit at the head of the establishment, can no longer ignore the problem. “So we need to listen more to the youth. We need to involve them more in a number of initiatives. AI and the ethics of AI are part of it,” she suggested.
“Let’s be honest, I’m only here for a few more years and speaking for myself. But the people who are, say, 15 and 18 years old today are the ones who have at least 50 years to go. If we are looking for long-term solutions and long-term commitments, they will be the ones who have to implement them, so they have to be included in this discussion,” Squicciarini said emphatically.
Dealing with data and algorithmic bias
While it sounds fair to develop a global AI policy, is there a framework or system to address biases in data and models and their consequences? “So your question is very important, because actually UNESCO cannot check every single algorithm,” Squicciarini said.
However, they are collaborating with researchers to develop an algorithm that checks other algorithms for bias, gender representativeness, identification, discrimination related to the outcome, and so on.
Squicciarini said the biases come from the person coding them and the data used to train the model. She said there are many ways codes can be customized and reviewed. “While UNESCO doesn’t have it, we really need an army of people to be able to do this on a global scale,” she added. However, she suggested that it relies on the company’s self-assessment at this point and that there must be accountability to remove data and algorithmic bias, which requires education and awareness work.