#FactCheck- AI-Generated Video Falsely Claims Iran Shot Down US F-35 Fighter Jet

Introduction

In September 2024, the Australian government announced the  Communications Legislation Amendment (Combatting Misinformation and Disinformation) Bill 2024 ( CLA Bill 2024 hereon), to provide new powers to the Australian Communications and Media Authority (ACMA), the statutory regulatory body for Australia's communications and media infrastructure, to combat online misinformation and disinformation. It proposed allowing the ACMA to hold digital platforms accountable for the “seriously harmful mis- and disinformation” being spread on their platforms and their response to it, while also balancing freedom of expression. However, the Bill was subsequently withdrawn, primarily over concerns regarding the possibility of censorship by the government. This development is reflective of the global contention on the balance between misinformation regulation and freedom of speech. 

Background and Key Features of the Bill

According to the BBC’s Global Minds Survey of 2023, nearly 73% of Australians struggled to identify fake news and AI-generated misinformation. There has been a substantial rise in misinformation on platforms like Facebook, Twitter, and TikTok since the COVID-19 pandemic, especially during major events like the bushfires of 2020 and the 2022 federal elections. The government’s campaign against misinformation was launched against this background, with the launch of The Australian Code of Practice on Disinformation and Misinformation in 2021. The main provisions of the CLA Bill, 2024 were: 

1. Core Transparency Obligations of Digital Media Platforms: Publishing current media literacy plans,  risk assessment reports, and policies or information on their approach to addressing mis- and disinformation. The ACMA would also be allowed to make additional rules regarding complaints and dispute-handling processes.
2. Information Gathering and Record-Keeping Powers: The ACMA would form rules allowing it to gather consistent information across platforms and publish it. However, it would not have been empowered to gather and publish user information except in limited circumstances.
3. Approving Codes and Making Standards: The ACMA would have powers to approve codes developed by the industry and make standards regarding reporting tools, links to authoritative information, support for fact-checking, and demonetisation of disinformation. This would make compliance mandatory for relevant sections of the industry. 
4. Parliamentary Oversight: The transparency obligations, codes approved and standards set by ACMA under the Bill would be subject to parliamentary scrutiny and disallowance. ACMA would be required to report to the Parliament annually. 
5. Freedom of Speech Protections: End-users would not be required to produce information for ACMA unless they are a person providing services to the platform, such as its employees or fact-checkers. Further,  it would not be allowed to call for removing content from platforms unless it involved inauthentic behavior such as bots. 
6. Penalties for Non-Compliance: ACMA would be required to employ a “graduated, proportionate and risk-based approach” to non-compliance and enforcement in the form of formal warnings, remedial directions, injunctions, or significant civil penalties as decided by the courts, subject to review by the Administrative Review Tribunal (ART). No criminal penalties would be imposed.

Key Concerns

• Inadequacy of Freedom of Speech Protections: The biggest contention on this Bill has been regarding the issue of possible censorship, particularly of alternative opinions that are crucial to the health of a democratic system. To protect the freedom of speech, the Bill defined mis- and disinformation, what constitutes “serious harm” (election interference, harming public health, etc.), and what would be excluded from its scope. However, reservations among the Opposition persisted due to the lack of a clear mechanism to protect divergent opinions from the purview of this Bill. 
• Efficacy of Regulatory Measures: Many argue that by allowing the digital platform industry to make its codes, this law lets it self-police. Big Tech companies have no incentive to curb misinformation effectively since their business models allow them to reap financial benefits from the rampant spread of misinformation. Unless there are financial non- or dis- incentives to curb misinformation, Big Tech is not likely to address the situation at war footing.  Thus, this law would run the risk of being toothless. Secondly, the Bill did not require platforms to report on the “prevalence of” false content which, along with other metrics,  is crucial for researchers and legislators to track the efficacy of the current misinformation-curbing practices employed by platforms.  
• Threat of Government Overreach:  The Bill sought to expand the ACMA’s compliance and enforcement powers concerning misinformation and disinformation on online communication platforms by giving it powers to form rules on information gathering, code registration, standard-making powers, and core transparency obligations. However, even though the ACMA as a regulatory authority is answerable to the Parliament, the Bill was unclear in defining limits to these powers. This raised concerns from civil society about potential government overreach in a domain filled with contextual ambiguities regarding information.

Conclusion

While the Communications Legislation Amendment (Combatting Misinformation and Disinformation) Bill sought to equip the ACMA with tools to hold digital platforms accountable and mitigate the harm caused by false information, its critique highlights the complexities of regulating such content without infringing on freedom of speech. Legislations and proposals regarding the matter all over the world are having to contend with this challenge. Globally, legislation and proposals addressing this issue face similar challenges, emphasizing the need for a continuous discourse at the intersection of platform accountability, regulatory restraint, and the protection of diverse viewpoints.

To regulate Big Tech effectively, governments can benefit from adopting a consultative, incremental, and cooperative approach, as exemplified by the European Union’s Digital Services Act 2023. Such a framework provides for a balanced response, fostering accountability while safeguarding democratic freedoms.

Resources

• https://www.infrastructure.gov.au/sites/default/files/documents/factsheet-misinformation-disinformation-bill.pdf 
• https://www.infrastructure.gov.au/have-your-say/new-acma-powers-combat-misinformation-and-disinformation 
• https://www.mi-3.com.au/07-02-2024/over-80-australians-feel-they-may-have-fallen-fake-news-says-bbc 
• https://www.hrlc.org.au/news/misinformation-inquiry 
• https://humanrights.gov.au/our-work/legal/submission/combatting-misinformation-and-disinformation-bill-2024 
• https://www.sbs.com.au/news/article/what-is-the-misinformation-bill-and-why-has-it-triggered-worries-about-freedom-of-speech/4n3ijebde 
• https://www.hrw.org/report/2023/06/14/no-internet-means-no-work-no-pay-no-food/internet-shutdowns-deny-access-basic#:~:text=The%20Telegraph%20Act%20allows%20authorities,preventing%20incitement%20to%20the%20commission 
• https://www.hrlc.org.au/submissions/2024/11/8/submission-combatting-misinformation?utm_medium=email&utm_campaign=Media%20Release%20Senate%20Committee%20to%20hear%20evidence%20calling%20for%20Albanese%20Government%20to%20regulate%20and%20hold%20big%20tech%20accountable%20for%20misinformation&utm_content=Media%20Release%20Senate%20Committee%20to%20hear%20evidence%20calling%20for%20Albanese%20Government%20to%20regulate%20and%20hold%20big%20tech%20accountable%20for%20misinformation+Preview+CID_31c6d7200ed9bd2f7f6f596ba2a8b1fb&utm_source=Email%20campaign&utm_term=Read%20the%20Human%20Rights%20Law%20Centres%20submission%20to%20the%20inquiry

‍

Artificial intelligence is revolutionizing industries such as healthcare to finance to influence the decisions that touch the lives of millions daily. However, there is a hidden danger associated with this power: unfair results of AI systems, reinforcement of social inequalities, and distrust of technology. One of the main causes of this issue is training data bias, which appears when the examples on which an AI model is trained are not representative or skewed. To deal with it successfully, this needs a combination of statistical methods, algorithmic design that is mindful of fairness, and robust governance over the AI lifecycle. This article discusses the origin of bias, the ways to reduce it, and the unique position of fairness-conscious algorithms.

‍

Why Bias in Training Data Matters

‍

The bias in AI occurs when the models mirror and reproduce the trends of inequality in the training data. When a dataset has a biased representation of a demographic group or includes historical biases, the model will be trained to make decisions in ways that will harm the group. This is a fact that has a practical implication: prejudiced AI may cause discrimination during the recruitment of employees, lending, and evaluation of criminal risks, as well as various other spheres of social life, thus compromising justice and equity. These problems are not only technical in nature but also require moral principles and a system of governance (E&ICTA).

Bias is not uniform. It may be based on the data itself, the algorithm design, or even the lack of diversity among developers. The bias in data occurs when data does not represent the real world. Algorithm bias may arise when design decisions inadvertently put one group at an unfair advantage over another. Both the interpretation of the model and data collection may be affected by human bias. (MDPI)

‍

Statistical Principles for Reducing Training Data Bias 

‍

Statistical principles are at the core of bias mitigation and they redefine the data-model interaction. These approaches are focused on data preparation, training process adjustment, and model output corrections in such a way that the notion of fairness becomes a quantifiable goal.

‍

Balancing Data Through Re-Sampling and Re-Weighting

‍

Among the aforementioned methods, a fair representation of all the relevant groups in the dataset is one way. This can be achieved by oversampling underrepresented groups and undersampling overrepresented groups. Oversampling gives greater weight to minority examples, whereas re-weighting gives greater weight to under-represented data points in training. The methods minimize the tendency of models to fit to salient patterns and improve coverage among vulnerable groups. (GeeksforGeeks)

‍

Feature Engineering and Data Transformation

‍

The other statistical technique is to convert data characteristics in such a way that sensitive characteristics have a lesser impact on the results. In one example, fair representation learning adjusts the data representation to discourage bias during the untraining of the model. The disparate impact remover adjust technique performs the adjustment of features of the model in such a way that the impact of sensitive features is reduced during learning. (GeeksforGeeks)

‍

Measuring Fairness With Metrics

‍

Statistical fairness measures are used to measure the effectiveness of a model in groups. 

‍

Fairness-Aware Algorithms Explained

‍

Fair algorithms do not simply detect bias. They incorporate fairness goals in model construction and run in three phases including pre-processing, in-processing, and post-processing.

‍

Pre-Processing Techniques

‍

Fairness-aware pre-processing deals with bias prior to the model consuming the information. This involves the following ways:

‍

• Rebalancing training data through sampling and re-weighting training data to address sample imbalances.
• Data augmentation to generate examples of underrepresented groups.
• Feature transformation removes or downplays the impact of sensitive attributes prior to the commencement of training. (IJMRSET)

‍

These methods can be used to guarantee that the model is trained on more balanced data and to reduce the chances of bias transfer between historical data.

‍

In-Processing Techniques

‍

The in-processing techniques alter the learning algorithm. These include:

‍

• Fairness constraints that penalize the model for making biased predictions during training.
• Adversarial debiasing, where a second model is used to ensure that sensitive attributes are not predicted by the learned representations.
• Fair representation learning that modifies internal model representations in favor of 

‍

Post-Processing Techniques

‍

Fairness may be enhanced after training by changing the model outputs. These strategies comprise:

‍

• Threshold adjustments to various groups to meet conditions of fairness, like equalized odds.
• Calibration techniques such that the estimated probabilities are fair indicators of the actual probabilities in groups. (GeeksforGeeks)

‍

Challenges 

‍

Mitigating bias is complex. The statistical bias minimization may at times come at the cost of the model accuracy, and there is a conflict between predictive performance and fairness. The definition of fairness itself is potentially a difficult task because various applications of fairness require various criteria, and various criteria can be conflicting. (MDPI)

Gaining varied and representative data is also a challenge that is experienced because of privacy issues, incomplete records, and a lack of resources. The auditing and reporting done on a continuous basis are needed so that mitigation processes are up to date, as models are continually updated. (E&ICTA)

‍

Why Fairness-Aware Development Matters

‍

The outcomes of the unfair treatment of some groups by AI systems are far-reaching. Discriminatory software in recruitment may support inequality in the workplace. Subjective credit rating may deprive deserving people of opportunities. Unbiased medical forecasts might result in the flawed allocation of medical resources. In both cases, prejudice contravenes the credibility and clouds the greater prospect of AI. (E&ICTA)

Algorithms that are fair and statistical mitigation plans provide a way to create not only powerful AI but also fair and trustworthy AI. They admit that the results of AI systems are social tools whose effects extend across society. Responsible development will necessitate sustained fairness quantification, model adjustment, and upholding human control.

‍

Conclusion 

‍

AI bias is not a technical malfunction. It is a mirror of real-world disparities in data and exaggerated by models. Statistical rigor, wise algorithm design, and readiness to address the trade-offs between fairness and performance are required to reduce training data bias. Fairness-conscious algorithms (which can be implemented in pre-processing, in-processing, or post-processing) are useful in delivering more fair results. As AI is taking part in the most crucial decisions, it is necessary to consider fairness at the beginning to have a system that serves the population in a responsible and fair manner.

‍

References

‍

• Understanding Bias in Artificial Intelligence: Challenges, Impacts, and Mitigation Strategies: E&ICTA, IITK
• Bias and Fairness in Artificial Intelligence: Methods and Mitigation Strategies: JRPS Shodh Sagar 
• Fairness and Bias in Artificial Intelligence: A Brief Survey of Sources, Impacts, and Mitigation Strategies: MDPI 
• Ensuring Fairness in Machine Learning Algorithms: GeeksforGeeks
  Bias and Fairness in Machine Learning Models: A Critical Examination of Ethical Implications: IJMRSET 
• Bias in AI Models: Origins, Impact, and Mitigation Strategies: Preprints
• Bias in Artificial Intelligence and Mitigation Strategies: TCS
• Survey on Machine Learning Biases and Mitigation Techniques: MDPI

‍

Introduction

Significantly, in March 2023, the Prevention of Money Laundering Act, 2002's regulations placed Virtual Digital Asset Service Providers functioning located under the purview of the Anti Money Laundering/Counter Financing of Terrorism (AML-CFT) scheme. An important step toward controlling VDA SP operations and guaranteeing adherence to Anti-Money Laundering and Combating the Financing of Terrorism (AML-CFT) regulations.

The significance of AML-CFT procedures

The AML-CFT framework's incorporation of Virtual Digital Asset Service Providers (VDA SPs) is essential for protecting the banking industry from illegal activities including the laundering of funds and counter-financing of terrorist attacks. These regulations become more crucial as the market for digital assets develops and becomes more well-known.

The practice of money laundering is hiding the source of the sum received illegally, thus it's critical to have strict policies in place to track down and stop these kinds of operations. Furthermore, funding for terrorism is a serious danger to international safety, hence stopping the flow of money to terrorist companies is a top concern for global officials.

The goal of policymakers' move to include VDA SPs in the AML-CFT architecture is to set up control and surveillance procedures that will guarantee these organisations' open and honest operations. This involves tracking transactions, flagging questionable activity, and conducting extensive customer investigations. Incorporating such procedures not only reduces the potential for financial crimes but also builds confidence and trust in the electronic asset market.

It is important to see the significance of AML-CFT procedures and the changes in the legal framework to reflect the evolving characteristics of digital currencies. These procedures are essential to preserving the reliability and safety of the wider banking system.

Notifications of Compliance Show Cause

Under Section 13 of the PMLA Act 2002, FIU IND sent adherence Show Cause Notices to nine offshore Virtual Digital Asset Service Providers (VDA SPs) as part of its dedication to upholding compliance with regulations. This affirmative step requires organisations to be scrutinised and attempted to bring them under inspection.

Governmental Response

The Director of FIU IND has addressed the Secretary of the Ministry of Electronics and Information Technology to take further measures due to the disregard of offshore firms. According to the notification, URLs connected to these organisations that operate in India in violation of the PML Act's requirements must be blocked. 

Mandatory Registration for VDA SPs

Virtual Digital Asset Service Providers (both onshore and offshore) who perform a range of operations, including the trading of digital goods for monetary currencies, the distribution of digital currency, and the management or preservation of electronic assets, are now obliged to register with FIU.

Range of Statutory Responsibilities

In accordance with the PML Act, VDA SPs are subject to several requirements, including documentation, disclosure, and other duties. One of their responsibilities is to register with the FIU IND. The primary focus is on guaranteeing that VDA SPs comply with AML-CFT protocols, hence enhancing the general reliability of the banking industry.

Difficulties with Offshore Compliance

There are many obstacles in guaranteeing that offshore organisations comply with Anti Money Laundering/Counter Financing of Terrorism (AML-CFT), chief amongst them being their unwillingness to undergo registration. Some overseas Virtual Digital Asset Service Providers (VDA SPs) have been reluctant to comply with the existing rules and regulations, even though they cater to a significant number of Indian users. There are several reasons for this hesitation, such as worries about heightened monitoring, the expense of compliance, and the apparent complexity of governmental processes. Regulatory organisations have taken steps to close the discrepancy between offshore businesses' real activities and the regulations they must follow. In addition to maintaining the trustworthiness of the economic system, resolving the issues with offshore adherence is essential for promoting confidence and openness in the market for electronic assets.

Conclusion

FIU IND has demonstrated its dedication to creating an effective regulatory framework for Virtual Digital Asset Service Providers through its recent measures. India hopes to fortify its countermeasures against money laundering and safeguard the financial well-being of its users by expanding the AML-CFT legislation to offshore firms. The continuous efforts to restrict the URLs of non-compliant companies show a proactive approach to stopping illicit activity and fostering a safe and law-abiding virtual asset ecosystem. The safety and soundness of the banking sector will be crucially maintained by laws and regulations as the digital world develops.

References

• https://pib.gov.in/PressReleasePage.aspx?PRID=1991372
• https://www.thehindubusinessline.com/books/reviews/business-economy/fiu-ind-issues-compliance-showcause-notices-to-nine-offshore-vda-sps/article67684613.ece
• https://business.outlookindia.com/news/fiu-issues-notice-to-9-offshore-crypto-platforms-writes-to-meity-for-blocking-of-urls