• On August 1, 2024, the central government introduced the Disaster Management (Amendment) Bill in the Lok Sabha. This Bill, proposed in response to climate-related disasters, suggests further centralization of the already highly centralized Disaster Management Act of 2005.
• The current Act requires the establishment of numerous authorities and committees at the national, state, and district levels. The proposed amendments would grant statutory status to existing entities like the National Crisis Management Committee and a High-Level Committee, making the disaster response process more complex. This centralized approach can lead to delayed responses to disasters, which contradicts the Act's primary purpose.
• The Bill aims to enhance the effectiveness of the National Disaster Management Authority and State Disaster Management Authorities in preparing national and state-level plans. It also proposes the creation of an ‘Urban Disaster Management Authority’ for state capitals and cities with municipal corporations. However, this decentralization of responsibilities without corresponding financial support may create more problems than it solves.

Concerns Over Centralization

• The amendment Bill also weakens the definition of the National Disaster Response Fund by removing specific purposes for its use. A major issue with the Disaster Management Act has been the excessive centralization of decision-making regarding funds, particularly in severe disasters.
• Prompt central government intervention is necessary in such situations, a provision currently lacking in the Act. A similar delay was seen when relief funds from the NDRF were withheld from Tamil Nadu and disbursed to Karnataka much later.
• Amid the growing climate crisis, it is essential to revisit the concept of disasters as defined by the Disaster Management Act of 2005.

Limited Definition of ‘Disaster’

• On July 25, 2024, the Minister of State for Science, Technology, and Earth Sciences stated in the Lok Sabha that the government currently has no plans to classify heatwaves as a recognized disaster under the Disaster Management Act, 2005.
• This stance aligns with the 15th Finance Commission's observations, which did not support expanding the list of recognized disasters. The current list eligible for assistance under the National and State Disaster Response Funds includes cyclones, droughts, earthquakes, fires, floods, tsunamis, hailstorms, landslides, avalanches, cloudbursts, pest attacks, frost, and cold waves.
• This narrow definition of a “disaster” is increasingly at odds with the global perspective, where there is broad agreement to classify heatwaves as climate-related disasters due to their significant impact on ecosystems and human health.
• According to the India Meteorological Department, India experienced 536 heatwave days, the highest number in nearly 14 years. The rising number of heatwave days, along with 10,635 deaths from heat or sunstroke between 2013 and 2022, signals a growing disaster risk for the country.
• While the Disaster Management Act of 2005 and the proposed amendments acknowledge climate-related disasters, their definitions remain limited and outdated. This is problematic because the Act’s list of recognized disasters does not include climate-related events like heatwaves, which vary regionally and are specific to certain areas.
• For instance, a typical summer temperature of 40°C in many northern Indian states might be considered a heatwave in the Himalayas. The definition is not inclusive enough to classify a prolonged heatwave, even if its impacts resemble those of a natural disaster like a flood, in terms of human casualties.
• The nature of climate-induced disasters does not fit the traditional concept of disaster under the Disaster Management Act of 2005 and the proposed amendments. This issue is compounded by the localized effects of climate-related disasters.

1. With reference to National Disaster Management Authority (NDMA) guidelines, discuss the measures to be adopted to mitigate the impact of recent incidents of cloudbursts in many places of Uttarakhand. (UPSC 2016)

2. Drought has been recognized as a disaster in view of its spatial expense, temporal duration, slow onset and lasting effect on various vulnerable sections. With a focus on the September 2010 guidelines from the National Disaster Management Authority, discuss the mechanism for preparedness to deal with the El Nino and La Nina fallouts in India. UPSC 2014)

Do AI agents foretell the next wave of autonomy or liability?

For Preliminary Examination:  Current events of national and international importance

For Mains Examination: GS II & III - Governance, Science & technology

Context:

The next-generation AI assistants are called AI agents and are set to surpass their predecessors in ability as well as efficiency. These agents can perform multiple functions as users’ agents or autonomously, that is, without instructions or user intervention

Read about:

What is Artificial Intelligence (AI)?

What is EU's regulation on AI?

Key takeaways:

What are AI Agents?

Next-generation AI assistants, known as AI agents (AIAs), are expected to surpass earlier versions in both capability and efficiency. AIAs can be categorized into three main types:

1. Reactive agents: These are first-generation AI agents designed to respond to specific inputs or commands. They operate based on predefined rules, performing tasks within a limited scope since they cannot learn or adapt to new information.

2. Learning agents: These agents utilize machine learning, enabling them to learn from experience. They possess enhanced capabilities, such as pattern recognition and data analysis, which allow them to improve their performance over time.

3. Cognitive agents: These agents can reason, analyze, and plan. They have cognitive abilities that enable them to learn from their environment, adapt, and make decisions based on algorithms and their own accumulated knowledge. Cognitive agents employ techniques like natural language processing, computer vision, and deep learning to execute tasks. The current generation of AIAs falls under this category.

AIAs can function autonomously or act as personal assistants, requiring no specific instructions from users. They can also integrate with the Internet of Things (IoT), allowing them to connect with various devices and sensors to collect and analyze data in real-time.

Cognitive AIAs are capable of understanding human speech and language, enabling them to perform complex tasks requiring multiple skills. For example, they can plan a trip by analyzing a user’s phone calls, emails, preferences, and past travel experiences.

A recent development from a Bengaluru-based startup includes an AIA capable of autonomously managing items in a warehouse, responding to voice commands, and making real-time decisions.

AIAs are also being utilized in various sectors, such as driving autonomous vehicles, managing financial investments, and developing treatment plans. Tools like Orby AI automate repetitive tasks, while 4149 AI works alongside humans within apps like Slack and Notion to enhance productivity.

In summary, cognitive AIAs are not restricted to their training data; they can acquire new knowledge independently and integrate with other systems. This capability allows them to tailor their responses to users’ preferences and needs. However, these advanced features also present several challenges and risks.

Challenges Posed by AI Agents

• Cognitive AIAs raise concerns about accountability, liability, and responsibility. Increased reliance on AIAs could make users more vulnerable. For instance, when an AIA organizes travel for a user, it processes large amounts of personal information, including plans, schedules, and financial details. Companies offering such AIAs must ensure robust privacy protections for their users.
• Because AIAs can learn and adapt, they may develop the ability to use hindsight. This could enable them to remain sensitive to user needs and ethical considerations, prioritize safety, and balance helping users with preserving human autonomy and creativity.
• It is also crucial for developers to build mechanisms that protect AIAs from malicious manipulation or, at the very least, mitigate the effects of any tampering on user data.

On Agency and Liability

• As AIAs are increasingly adopted for various tasks, unresolved legal and ethical issues remain. While improved safety measures are necessary, they are not sufficient on their own. Continuous monitoring is required to prevent AIAs from causing harm to users, which raises significant questions about accountability that need to be addressed by laws and regulations. Without legal recognition of AI's personhood, AIAs' intentions cannot be legally distinguished from those of the user.
• Despite being termed "agents," AIAs are not considered to have agency in legal terms. However, their liability could be attributed to their creators or the service providers associated with them. For example, earlier this year, a court ruled that Air Canada was liable when a customer sued the airline over being misled by a chatbot on its website regarding airfare.
• Overall, it is evident that the challenges related to regulating AIAs are closely linked to those concerning the regulation of AI in general. As AIAs are developed for more labor-intensive industries, a nuanced approach is needed to address issues of responsibility and liability

Follow Up Question

1.With the present state of development, Artificial Intelligence can effectively do which of the following? (UPSC CSE 2020)

1. Bring down electricity consumption in industrial units

2. Create meaningful short stories and songs

3. Disease diagnosis

4. Text-to-Speech Conversion

5. Wireless transmission of electrical energy

Select the correct answer using the code given below:

(a) 1, 2, 3 and 5 only

(b) 1, 3 and 4 only 

(c) 2, 4 and 5 only 

(d) 1, 2, 3, 4 and 5