Artificial intelligence (AI) is evolving at a rapid pace, posing ethical challenges in addition to efficiencies in fields including healthcare, education, finance, and criminal justice. Apprehensions about prejudice, its impact on human decision-making, and its potential for propaganda can be raised. The algorithmic bias impedes the justice and fairness of marginalised people in vital domains like employment and healthcare. Although AI might be efficient, improper use of the technology can be a potential menace, highlighting the need for careful consideration and expert disposition. The possible exploitation of AI for propaganda demands robust content verification mechanisms and public awareness to combat misinformation effectively, safeguarding societal stability and democratic processes. These challenges are evident across various sectors, emphasizing the importance of a balanced approach to maximize the benefits of AI while mitigating its negative impacts.

Risk of Bias

Among the critical concerns surrounding AI is the inherent risk of algorithmic bias. This bias becomes apparent when AI algorithms are trained on data sets that are either non-representative or skewed, culminating in outcomes that are unjust or discriminatory. The study on biases in AI-powered systems affecting marginalized communities in job recruitment processes, highlighted in the article “Addressing bias in big data and AI for health care: A call for open science,” discusses the disparities in data sources and patient populations within the landscape of AI in clinical medicine. It emphasizes the sources of bias in AI systems that perpetuate healthcare disparities. Also, the article “Humans inherit artificial intelligence biases” delves into the impact of biases in AI systems on social inequalities and marginalised communities, with specific examples from India. It explores how AI-based technologies in India have exhibited biases related to race, gender, and caste, leading to unfair treatment, and hindering opportunities for marginalized individuals.

The implications of such biases are particularly alarming in contexts that directly influence the lives and future of individuals, such as loan approvals, job recruitment, and the criminal justice system. The propagation of existing biases through AI not only deepens social inequalities but also inflicts harm on marginalised communities. A concerted effort is thus required to ensure the development of AI systems with diverse, inclusive, and representative data, coupled with rigorous testing and validation to identify and rectify biases.

Impact on Decision-Making

The influence of AI on human decision-making is evident in algorithms in streaming services, social media, and online retail. However, the “filter bubble” effect restricts the diversity of thought, potentially impairing critical thinking. Ensuring transparency and user control is crucial. AI is increasingly being utilized in sectors like healthcare, finance, and education for various purposes, including disease diagnosis, fraud detection, market trend prediction, and personalized learning experiences.

However, this use in decision-making also raises concerns about the potential for bias and discrimination. For instance, in 2019, one study found that AI systems used by Indian banks to make lending decisions showed a bias against women and low-income groups. This highlights the importance of ensuring algorithmic transparency and accountability to prevent such biases from perpetuating in decision-making processes.

Application Oversight

The dependence on AI for various challenges has led to inefficiencies and resource wastage. This highlights the need for critical evaluation and validation of AI applications in specific domains to ensure they offer value and enhancement over existing methodologies.

Misapplication of AI in healthcare instead of diagnosing diseases and predicting patient outcomes can lead to biased diagnoses and predictions, potentially harming patients. In finance, AI can detect fraud and predict market trends, but improper validation can cause false alarms and financial losses.

To counteract these issues, it is crucial to ensure that AI systems are developed and deployed with proper expertise and understanding. This includes ensuring that the data used to train the AI system is representative and diverse, validating the system’s performance, and continuously monitoring its performance to identify and address any issues. Additionally, it is essential to educate stakeholders on the limitations and potential risks of AI applications to ensure that they are used responsibly and effectively.

Propaganda Tool

AI’s potential for propaganda and misinformation threatens societal stability and democratic processes. Addressing this requires robust verification mechanisms, transparency in AI-generated content deployment, and increased public awareness of AI’s capabilities and limitations.

The rise of AI-generated misinformation in India has become a pressing concern. Noted neuroscientist Mauktik Kulkarni highlighted the mushrooming of fake news sites using Artificial Intelligence, with a staggering 1000% increase in AI-generated misinformation. These fake news sites have been observed spreading falsehoods, including morphing videos of popular shows like “Kaun Banega Crorepati” to push political agendas and creating AI-generated speeches of deceased leaders like Swami Vivekananda.

The World Economic Forum’s Global Risk Report 2024 ranked India as the top country facing the crisis of AI-generated misinformation ahead of its parliamentary elections. The spread of deep fake videos and AI-driven fake news has raised concerns about influencing voters by targeting their existing beliefs. Misinformation in India often revolves around political and religious motivations, deepening social divides and polarizing communities.

To combat the spread of AI-generated fake news, fact-checking agencies in India face challenges due to the high cost of tools for identifying AI-generated content compared to those for creating them. The lack of financial incentives for detection hinders efforts to debunk online misinformation effectively. Additionally, political parties’ IT cells are identified as key sources of AI-generated fake news, complicating India’s misinformation landscape.

Conclusion

The rapid evolution of AI presents a complex area where stakeholders, including the government, technocrats, and end users, play crucial roles in shaping its trajectory. Government must establish regulatory frameworks to ensure the ethical development and deployment of AI, promoting transparency, accountability, and fairness. Technocrats must design algorithms prioritizing fairness and inclusivity, conduct risk assessments and advocate for ethical AI practices. End users, by interacting with AI technologies and providing real value feedback can contribute to shaping AI outcomes.

(Author is PhD Student, Pondicherry University. Feedback: [email protected])