Professor Toby Walsh, an expert in artificial intelligence (AI) at the University of New South Wales, Australia, talks to Kuensel’s Rinzin Wangchuk on the sidelines of the Drukyul Literature Festival. Excerpts. 

What are the potential dangers of AI in various domains, such as cybersecurity and job displacement, and how can these risks be mitigated?

I’ve written a whole book, my most recent book, which is trying to answer the many dimensions of that question. I don’t like to talk about the risks without also pointing out to people the many benefits. The reason that we’re building the technology is because of the way it’s going to improve our health, and our wealth, and hopefully allow us to live more sustainable lives.

AI is a technology that will touch every aspect of our lives, and so the risks also touch many aspects of our lives, from the displacement of jobs to misinformation that may disrupt elections. We’re already seeing some of that. To greater concerns that some people have, existential concerns that people have, I don’t think there is one solution to all of these challenges. There’s a combination of things.

There are technical solutions. I spend a lot of my time working, trying to develop AI in a way that is addressing some of the concerns. There are regulatory solutions. We are increasingly waking up to the idea that there are things that need to be better regulated.

Misinformation may require greater regulation. And then there are educational and societal solutions, which is that there’s also some responsibility upon the users of the technology to use it in responsible ways.

So I think a combination of all those factors will ensure that AI is a very large net positive. The expectation is that it will grow our wealth immensely and help us tackle many of the wicked problems we face from the climate emergency to increasing inequality to fractured political discourse. Responsible and thoughtful use of AI can lead to significant positive impacts. 

How can AI be ethically developed and deployed to ensure it aligns with human values and does not perpetuate biases or discriminatory practices?

In addressing the ethical development and deployment of AI to align with human values and avoid perpetuating biases, it’s crucial to acknowledge that while AI technologies themselves may not carry inherent biases, they can inadvertently perpetuate historical biases present in the data they are trained on. 

To mitigate this, technical solutions can be employed to reduce bias, along with regulatory measures to ensure fair and just systems. Users and consumers also play a role by being discerning and understanding the challenges associated with biased AI.

AI offers the opportunity to make more evidence-based and fairer decisions compared to humans, who often harbour subconscious and conscious biases. However, achieving this requires careful and diligent efforts. It’s important to recognize that even with the best methods and algorithms, there are instances where completely unbiased answers are unattainable. For example, in hiring processes, selecting a few candidates from a larger pool inherently introduces bias, such as favouring candidates with higher grades, which may be influenced by socio-political factors.

Ultimately, society must determine the values and norms it wishes to uphold in AI decision-making. For instance, ensuring gender-neutral outcomes is a priority while determining acceptable biases related to age is a matter of societal choice. Striking a balance between unbiased decision-making and practicality is complex, but open dialogue and careful consideration of ethical implications are essential to ensuring AI’s alignment with human values.

What are the potential long-term societal impacts of widespread AI adoption, and how can we prepare for and manage these changes effectively?

I think widespread AI adoption has immense potential for positive impacts, particularly in education and healthcare, offering personalized and high-quality services to all. AI could potentially revolutionize disease treatment and resource management, leading to a more sustainable world. 

However, there are risks and challenges associated with AI. It can be misused for surveillance, exacerbate inequality, disrupt democratic processes, and spread misinformation. To effectively manage these changes, we must implement robust regulations, promote transparency and accountability in AI systems, educate the public, and foster collaboration between various stakeholders. While AI offers great benefits, careful management is crucial to ensure it aligns with societal values and avoids detrimental consequences.

Can you explain some real-world examples where AI has been beneficial in improving human lives, healthcare, or the


AI has demonstrated significant benefits in improving human lives, healthcare, and the environment. In healthcare, AI has been instrumental in addressing the challenge of antibiotic resistance. By using AI to discover new antibiotics more efficiently, we have the potential to combat drug-resistant bacteria and save lives. AI has already identified two promising antibiotics that are currently in clinical trials.

Regarding the environment, AI has shown its prowess in optimizing supply chains and reducing carbon emissions. Collaborating with large multinational companies, AI has successfully optimized transportation routes, leading to a 15 percent reduction in carbon dioxide emissions. This significant environmental impact is not a one-time occurrence; it results in sustained yearly reductions.

AI’s ability to process vast amounts of data and make complex decisions has proven invaluable in these real-world examples. From advancing healthcare treatments to mitigating environmental impacts, AI has the potential to revolutionize various fields and positively impact human lives and the planet. Continued efforts in AI research and application hold the promise of even more transformative solutions in the future.

Given the resource constraints of a small country like Bhutan, what are the most critical risks and dangers associated with adopting AI technologies, and how can these risks be addressed?

I think the biggest risk is that you’re a very small pebble in a very large sea. The most critical risks for a small country like Bhutan in adopting AI technologies are the potential loss of value to companies outside the country and becoming a mere profit centre for them. It’s essential for Bhutan to be a quick adopter of AI, leveraging its unique resources such as a robust healthcare system, education system, and valuable data on the environment and forestry.

While Bhutan may not have the resources of tech giants, it can focus on utilizing its data for the benefit of its people and potentially the entire planet. For instance, like Iceland, Bhutan can use its isolated population and excellent health records to uncover valuable knowledge about the human genome and share that knowledge with the world.

The key is to innovate and find ways to harness AI to address local challenges and contribute positively to global knowledge and progress. By focusing on leveraging its unique data and resources, Bhutan can ensure that AI adoption brings value and benefits to its people, maintaining its identity as a kind and beautiful country while making significant contributions to the broader world.

Considering the limited availability of skilled AI professionals in Bhutan, what approaches should we take to develop local AI expertise and build capacity for responsible AI development and deployment?

I think there’s a lot to be said for improving the digital literacy of your citizens and developing local AI expertise and building capacity for responsible AI development in Bhutan can begin by focusing on improving digital literacy among citizens. 

An excellent example to follow is Finland, which aims to make 10 percent of its population, approximately 300,000 people, AI literate through a freely-accessible online course. This course teaches people about AI’s benefits, limitations, risks, and opportunities without requiring programming skills.

By implementing a similar initiative in Bhutan, with the support and involvement of the government and large companies, a significant fraction of the population can become AI literate. This ambition is achievable for a small country like Bhutan, with a population of about three-quarters of a million people.

I think enhancing AI literacy in the population will foster innovation and responsible AI use in Bhutan. Making citizens aware of AI’s potential and ethical considerations will enable them to contribute to AI development and deployment responsibly. By becoming early adopters of AI in a responsible manner, Bhutan can leverage the power of these technologies to address local challenges and make positive contributions to its society and the global community.