Leading AI scientists are talking about various existential risks associated with AI. It is essential for all of us to know the pre-emptive corrective measures that can be taken. Pausing the AI developments is not a practical idea and does not serve the purpose as it is impossible to track if someone is doing it in their backyard.

Scientist community itself has to further accelerate the research to make it safe from the adversaries to avoid any misuse. Guardrails and pre-emptive protocols need to be developed so that the applications are developed only for good of humanity. Following are important areas to be considered for AI to be good for the planet.

Mitigating these AI-related risks is a complex challenge that requires a coordinated, global effort. Different stakeholders – including governments, organizations, and individuals – all have a role to play. Here are some strategies to consider:

Develop comprehensive laws, regulations, and Oversight: Governments should establish clear regulations and oversight mechanisms for the development and use of AI. Governments can pass laws that clearly define the acceptable uses of AI, data privacy standards, and penalties for misuse. This could include things like requirements for transparency and explainability and measures to prevent misuse of AI by rogue actors.  Governments can establish regulatory bodies to oversee AI development and deployment, like the role of the FDA for drugs or the FAA for aviation. These bodies can conduct audits, issue licenses for certain uses of AI, and ensure compliance with regulations.

International treaties and AI Ethics Charter: Like treaties that govern nuclear weapons or climate change, nations can collectively agree to limit the use of certain AI technologies, such as autonomous weapons. The global community can also agree on AI ethics principles and cooperate on AI safety research. UN should adopt ethical guidelines for AI use. These could cover areas like data privacy, fairness, transparency, and accountability. AI should be designed and used in a way that respects human rights and promotes societal well-being.

International Cooperation, Collaboration and Development: Given the global nature of AI and its impacts, international cooperation is crucial. This could involve sharing best practices, coordinating on regulation, and working together to address global challenges like misinformation or cyber warfare. AI developers should commit to responsible practices, such as transparency in their AI systems’ capabilities and limitations, not contributing to harmful applications of AI, and prioritizing long-term safety in their research. Developers should also share their decision-making processes, data sources, and error rates. 

Human-in-the-loop and public participation: Keeping humans in the decision-making loop can help ensure that AI systems don’t make decisions that are unethical or harmful. This might involve using AI as a tool to aid human decision-making, rather than replacing it entirely. The public should be involved in decision-making processes about how AI is used in society. This could be through public consultations, citizen assemblies, or other forms of participatory decision-making. 

Research and Training: Ongoing research and development is needed to address the technical challenges associated with AI, such as ensuring AI systems behave as intended, can be controlled by humans, and don’t amplify biases. Individuals need to be educated about the potential risks and benefits of AI, and how to use it responsibly. This could be part of formal education, workplace training, or public awareness campaigns.

Handling Rogue organizations and Countries:  a) Sanctions and penalties: International bodies can impose sanctions or penalties on countries or organizations that violate agreed-upon AI ethics principles or regulations. b) Diplomacy and negotiation: Diplomatic efforts can be made to encourage rogue nations to abide by international norms and agreements. C) Cyber defense strategies: Countries need robust cyber defence strategies to protect against misuse of AI by rogue actors. This can include developing AI systems that can detect and defend against AI-powered cyber-attacks. d) International cooperation: It’s crucial for nations to work together to manage the risks associated with AI. This can involve sharing intelligence about rogue actors, coordinating responses, and working together on AI safety research.

However, it’s important to note that while these strategies can help, they are not guaranteed to be fully effective. The global community needs to stay vigilant and adaptable as the AI landscape continues to evolve.

This article first appeared in the TOI