The public choice approach to public administration started in the 1960s. Vincent Ostrom suggested using economic methods and behavioral assumptions to analyze administration. Thus, public choice approach emphasises the following :
- Consumer Oriented Approach : It lays emphasis on the element of choice, with the citizen in the role of consumer. It sees citizens as rational decision-makers who act in their self-interest. It stresses aligning government actions with citizen preferences and values.
- Institutional pluralism, that is, plurality of agencies to promote consumers’ preferences
- Democratic Decentralization: Dispersal of decision making authority to stimulate healthy and democratic competition among government agencies.
- Popular participation in administration and Debureaucratization.
- Political approach to public administration by locating public administration within the domain of politics
The public-choice approach is opposed to: (i) Centralized bureaucratic system of administration; (ii) Centralized system of planning and policy-making; (iii) Dichotomy between politics and administration; (iv) Non-committal bureaucracy; and (v) Monocentric, hierarchical bureaucratic administration.
Deepfakes are digitally synthesized media created using advanced deep learning techniques like Generative Adversarial Networks (GAN), to produce realistic but entirely fabricated visual and audio content.For example, a viral video featuring an actress in a swimsuit or the Prime Minister performing Garba.
The incorporation of hyper-realistic digital falsification presents the following challenges:
- Misinformation and Disinformation: Spread false information, manipulate public opinion, damage credibility. (E.g., deepfake of Ukrainian President Zelenskyy).
- Impersonation and Identity Theft: Lead to financial fraud.
- Privacy Violations: Use personal photos or videos without consent, lead to privacy violations.
- Legal and Ethical Concerns of AI: Raise questions about legality and ethics, call for regulation and accountability.
- Weaponization Against Women: Primarily pornographic, causing psychological trauma and social repercussions.
- Threats to National Security: Undermine public safety, create chaos, stir anti-state sentiments.
- Undermine Trust in Democratic Institutions: Manipulate people, sabotage elections. (E.g., viral video of Manoj Tiwari).
- Declining Trust in Traditional Media: Causes short-term and long-term harm.
To ensure the safety of Digital Nagrik, MeitY recently sent social media platforms advisories on deepfakes under IT Rules 2021.
Machine learning, a subset of artificial intelligence, enables computers to learn from data, identify patterns, and make decisions with minimal human intervention, automating the process of analytical model building.
- How it works:
- Decision Process: Applying the trained model to make predictions.
- Error Function: Evaluating model performance by measuring prediction accuracy.
- Model Optimization: Improving model accuracy by adjusting weightage to parameters to minimize errors.
Application- (2-4 bahut hai vese to)
- Speech Recognition: Eg-Â Siri, Alexa, etc.
- Customer Service: Eg- Slush, Maya Chatbots, etc.
- Computer Vision: Eg-Â Google Translate, Facebook 3D Photo, Faceapp, etc.
- Recommendation Engines:Eg-Â online ads, Spotify for recommending songs, etc.
- Automated Stock Trading:Eg-Â NinjaTrader, etc.
- Fraud Detection: Eg-Â Credit card fraud detection, POS fraud detection, etc.
Popular Methods: supervised learning (trained on labelled data), unsupervised learning (works with unlabeled data to find patterns), and reinforcement learning (agents learn through feedback).
- NITI Aayog defines AI as: AI is a constellation of technologies that enable machines to act with higher levels of intelligence and emulate the human capabilities of sense, comprehend and act.Â
- AI technologies offer various opportunities and benefits, they also present challenges and potential risks.
Opportunities | Challenges |
Eg. 3Nethra : Screening for common eye problems
|
|
Feature | Supervised Learning | Unsupervised Learning |
Data Type | Uses labeled data (input with known output). | Uses unlabeled data (data without known outcomes). |
Objective | Learn patterns to predict or classify future data (new, unseen data). | Discover hidden patterns or groupings in data. |
Example | Identifying cats in labeled photos (“cat” or “not cat”). | Grouping similar shopping items without labels. |
Common Models | Linear Regression, Logistic Regression, Support Vector Machines (SVM), Decision Trees, K-Nearest Neighbors (KNN). | K-Means Clustering, Hierarchical Clustering, Principal Component Analysis (PCA), Autoencoders etc. |
Applications |
|
|
Machine learning is a subset of artificial intelligence that enables computers to learn from data, identify patterns, and make decisions with minimal human intervention, thus automating the process of analytical model building.
Eg.
- Gmail automatically classifying emails as ‘Spam’ and ‘Not Spam;Â
- Flipkart or Amazon recommending you to buy products of your choice.Â
Aspect | Machine Learning | Traditional Programming |
Process | input data and the output data are fed to an algorithm (Machine learning algorithm) to create a program.  | refers to any manually created program which uses input data, runs on a computer and produces the output  |
Data-driven approach | Learns from data to adapt and improve over time. | Relies on predefined rules and logic. |
Generalization | Generalizes patterns to make predictions on new, unseen data. | Creates specific rules for each scenario. |
Automation | Automates learning and decision-making processes based on data. | Requires manual updates to rules and logic. |
Scalability | Can handle large and complex datasets effectively. | May struggle to scale for large data volumes. |
Iterative improvement | Models can be trained iteratively to refine performance over time. | Code is rarely updated unless necessary. |
