AI brings the advancements of economies and societies, but also a list of novel, ethical and social challenges.

Artificial Intelligence (AI) technology is quickly becoming a potential disrupter and essential enabler for almost all industries. Ultimately, organisations are also looking for ways to deploy AI solutions in their working system. What stagnates the process is the lack of assurance that AI is trustworthy and won't defy human at any stage.

The portrayal of futuristic movies has given mankind an impression that AI is scary after a point of time. Such dystopian fantasies about AI are widely discussed by experts and researchers as well. Even though such horrible AIs are a long way from where we are now, the world can stop it from happening by putting a trustworthy AI framework that limits and sets guidelines for innovations. AI brings the advancements of economies and societies, but also a list of novel, ethical and social challenges. From a business perspective, the potential consequences of AI include everything from the lawsuit to regulatory fines, angry consumers, embarrassment, reputation damage and destruction of shareholders. Trustworthy AI (TAI) bases on the idea that trust builds the foundation of societies, economies, and sustainable development so that individuals and organisations realize the full potential of AI. As AI is already disrupting people's modern life, it is the need of the hour to contain it in a safe box and leverage growth with restriction.

Trustworthy AI framework

AI-powered robots are replacing humans in many stances. Experts are worried that this will stop the routine as people will start losing their jobs to AI. Henceforth, the possible consequences of AI can still pose significant threat to the human race. For example, Goldman Sachs replaced over 600 trades with AI-powered robots. Besides, many companies are embracing virtual assistants, chatbot services, data analytics, etc to handle human jobs. The predictions for the future portray a far worse picture. Experts anticipate that AI-powered robots will be more intelligent than humans by 2045, which is quite scary. Henceforth, it is the right time to set a framework based on the following dimensions.

Fair and unbiased

Bias is a threat to humankind. The earth houses people of different colours, religions, origins, geographical backgrounds, etc. Humans have predominantly taken over other humans for a long time in the discrimination grounds, which was more than terrible. Since the disruption era is already here and AI is designated to be the future of modern technology, trustworthy AI with fair sense and unbiased nature should be designed to protect mankind from further discriminations. Even though AI takes data from humans to train its system, the framework should limit feeding AI modules with such biased data.

Explainability and transparency

AI makes crucial decisions on its own without human interference based on machine learning algorithms and big data. Unfortunately, this monopoly action stops people from knowing why the AI model took such a decision. The scenario gets worse as most AI systems are not powered with an explainable system that validates its reason behind decision-making. Henceforth, developers need to build explainable AI that reasons its motives behind every action.

Robust and reliable

AI usage is already on a spike. However, to reach more adoption, AI must be as robust and reliable as the traditional systems, processes, and people it is augmenting or replacing. For AI to be trustworthy, it must be available when it is supposed to be and must generate consistent and reliable outputs by performing tasks properly in less than ideal conditions and when encountering unexpected situations and data.

European Union guidelines for trustworthy AI

Most of the European Union nations like France, Germany and Italy are some of the frontrunners of AI. The commission is known to take unified steps to control everything. It has performed no less when it comes to tackling technology-related challenges. The EU keeps a tab on social media giants and regulates them at times like the recent proposal of two laws: the Digital Services Act and Digital Markets Act to stop big tech giants from acquiring small businesses. The commission has presented an Ethical Guidelines for Trustworthy Artificial Intelligence in 2019. Some of the key takeaways from the framework are,

• Human agency and oversight- AI systems should empower human beings, allowing them to make informed decisions and fostering their fundamental rights. Besides, proper oversight mechanisms should also be ensured.

• Technical robustness and safety- AI systems need to be resilient and secure; at the same time have a substitute plan if they fall behind. AI should also be accurate, reliable and reproducible.

• Privacy and data governance- Adequate data mechanism should be ensured taking into account the quality and integrity of the data with concern to legitimised access to data.

• Transparency- AI mechanism should be very transparent that their decisions are explained in a manner adapted to the stakeholder concerned.

• Diversity, non-discrimination and fairness- Unfair biases should be avoided as it could have multiple negative implications from the marginalised of vulnerable groups to the exacerbation of prejudice and discrimination.

OECD principles on Artificial Intelligence (AI)

OECD and partner countries adopted the first set of intergovernmental policy guidelines on AI in 2019. The 36 OECD member countries along with Argentina, Brazil, Colombia, Costa Rica, Peru and Romania signed up the OECD Principles on Artificial Intelligence to regulate technology growth. Some of the key recommendations that the commission requested to the governments are,

• Facilitate public and private investment in research & development to spur innovation in trustworthy AI.

• Foster accessible AI ecosystems with digital infrastructure and technologies, and mechanism to share data and knowledge.

• Create a policy environment that will open the way to deployment of trustworthy AI systems.

• Equip people with the skills for AI and support workers to ensure a fair transition.

• Co-operate across order and sectors to share information, develop standards and word towards responsible stewardship of AI.