Synthesizing Human Expert Opinion to Assess News towards News

Authenticity

S. Sangeetha

, S. Padmapriya

, J. Sudha

, R. Sudha

, A. Vivekanandhan

and R. Valampuranayaki

Department of IT, E.G.S. Pillay Engineering College, Nagapattinam, Tamil Nadu, India

Department of CSE, A.V.C College of Engineering, Mannampandal, Tamil Nadu, India

Department of IT, A.V.C College of Engineering, Mannampandal, Tamil Nadu, India

Keywords: Media Manipulation, Fake News Impact, Socio‑Political Influence News Authenticity, Information

Verification, Internet News Challenges.

Abstract: Socio-Political impact on news dissemination, the media is flooded with News with biased headlines to

generate a substantial profit and sometimes due to some socio-political influence it becomes grey area and is

widely use to influence it. As a result, there are several issues that have come to the forefront. such as the

socio-political impact of news propagation, the declining independence of news media, and a straightforward

news assessment system have become increasingly relevant issues due to the rising use of internet news.

Political polarization, motivated reasoning, and social media algorithms are the cause of fake news. Fake

news, no matter the extent of impact, is definitely damaging. The essence is that the news, we read or receive,

shall have good sources and the information shared in the news shall be authentic. This study tries to find and

verify the fake news. The Natural Language Processing (NLP) model is trained iteratively using the

incremental data to achieve the desired output before being utilized for fake news detection.

1 INTRODUCTION

News media in modern news media, as a top for

public, are the basis for of information access and

discussion. But between the grand mission of

enlightening the masses, a darker undercurrent of

news fake news, sensationalism, and This reporting,

often referred to as agenda driven or policy driven

reporting, has emerged as a real threat to the integrity

of public discourse and democratic processes. So,

what I really want to do here is to investigate the

complex entanglement of issues at play in the ever-

evolving media with a particular emphasis on the

pernicious impact of biased headlines, money-driven

editorial choices, and socio-political pressures on

journalistic standards.

Located at the heart of this debate is the problem

of fake news, a misleading trend exacerbated in the

digital age, when information spreads quickly and

widely. Driven by political motives, profit or raw

opportunism, the purveyors of fake news-savvy

hoaxes take advantage of the failings of online

platforms, fostering narratives that substitute reality

with divisive fare that erodes confidence in traditional

institutions. Set against this context, fears about the

erosion of press freedoms, the proliferation of echo

chambers fuelled by social media algorithms and the

unleashing of disinformation as a weapon of political

warfare, have reignited calls to action and re-enter

discourse with urgency.

2 RELATED WORK

In (M. A. Al Mumin et al. 2019)- Detecting fake news

using a Machine Learning Approach is a more

detailed machine learning approach which is focused

on detecting the Fake news in online media. In this

paper, the authors introduce a new approach for

accurately detecting and classifying fake news

articles based on the framework of natural language

processing, social network analysis, and supervised

learning algorithms. This framework extracts textual

features from news articles (including linguistic

patterns, sentiment analysis, and credibility

indicators) and integrates them with social network

features (such as source reputation and dissemination

patterns) to improve classification performance. As

Sangeetha, S., Padmapriya, S., Sudha, J., Sudha, R., Vivekanandhan, A. and Valampuranayaki, R.

Synthesizing Human Expert Opinion to Assess News towards News Authenticity.

DOI: 10.5220/0013876100004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 2, pages

17-20

ISBN: 978-989-758-777-1

the authors show, their approach performs this

detection task on a large dataset with a number of

actual news articles, and they achieve promising

accuracy and efficiency results. In this study, we

propose an integrated framework comprising a

combination of state-of-the-art machine learning

algorithms and model-friendly linguistic and social

network features to detect and differentiate between

the genuine news and its false counterparts.

For example, in (B. Uddin et al. 2021) - Spread of

Misinformation is the authors use a large dataset of

articles and social media shares to analyse whether

patterns exist in the dissemination of

misinformation. They also find that misinformation

spreads more rapidly and reaches a larger audience

than does fact, in part because misinformation is

new, emotionally compelling, and fits the social

network structure. University of California, Berkeley

Fake News Spread: The Role of Social Media

Misinformation has become a standard term in this

paper due to the mechanisms in fake news spread. By

examining thousands of news articles and millions of

social media shares, the authors pinpoint factors and

patterns that contribute to the virality of

misinformation. As a foundational source, this study

shall improve the understanding of the dissemination

of fake news which shall pave the way to future

research in detection and prevention of

misinformation.

Reference (S. Nath and S. Ray 2022)- Automation

Detection of Fake News. The authors discuss

previous methodologies such as content-based

analysis, social network analysis, and machine

learning algorithms, comparing their pros and cons. It

discusses the challenges in fake news detection,

including dataset bias, algorithmic bias, and

adversarial attacks and provides directions for future

work. This paper presents them with detailed

information on advanced approaches to combat fake

news as well as the foundation for both researchers

and practitioners in the area. It provides a detailed

overview of automated Methods for detecting and

analyzing misinformation, thus serving as an

important reference to researchers and practitioners

in the area of misinformation detection. A systematic

literature review of existing methods is made,

emphasizing their pros, cons, and applicability in a

given context.

Media Literacy Intervention for Misinformation

in reference (M. Kowsher et al. 2021) - The authors

discuss various educational programs, initiatives and

interventions that attempted to foster critical thinking

skills, media competence. and the traditions of

individual and group responsible information

consumption. They isolate common attributes

associated with successful media literacy programs,

including interactive instructional approaches,

customized course content, and techniques for

fostering community involvement. (M. A. Al Mumin

et al. 2019) This review comes with

recommendations and describes effective strategies

to combat misinformation, which could be used by

educators, media organizations and policy-makers.

Drawing from a study of a dozen media literacy

programs, the authors outline key elements that

increase the likelihood of successful media literacy

campaigns and recommend data-supported strategies

for creating effective programs. The review provides

a field guide for how to correctly respond to the

dangers of misinformation and the importance of

critical thinking skills and responsible information

consumption habits in the digital age.

3 PROPOSED METHODOLOGY

These authors explore educational programs or

initiatives or interventions that promote critical

thinking skills and digital literacy, and habits of

responsible information consumption for individuals

and communities. They analyze the main elements

that underscore just how successful media literacy

interventions are, including interactive learning

styles, context-specific programs, and community-

facing initiatives. The review highlights important

takeaways on how to effectively combat

misinformation and provides actionable

recommendations for educators, policymakers and

media organizations. Through systematic review of

academic literature associated with a diverse array of

educational programs and initiatives, these authors

illuminate important elements that promote the

efficacy of media literacy interventions and provide

concrete recommendations for developing 588

evidence-based interventions. It's an invaluable guide

in navigating the challenges posed by

misinformation and on how to promote critical

thinking and responsible information consumption in

the digital age.

Google Cloud computer for News Authentication

Application consists of modules of News

Authentication Cloud server client creation Reverse

news classification Publish Authenticated news

Cloud Server Creation – During cloud server

creation authorization the module authenticates to the

selected cloud provider using secure credentials,

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

establishing a trustworthy connection for the next

operations. You are read well new server instances as

instance type so or geographical zone.

Moreover, it also supports features such as

automatic scaling, where the users can easily adjust

server capacity in real-time based on demand

patterns. It also features strong error handling to

elegantly handle exceptions and minimize operation

disruption. In addition, the module provides

comprehensive feedback on the creation of servers,

providing information on real-time status updates and

resource allotment.

02 data monitoring — In this module the data

posted as news will be monitored. As well as News

extracted from general data. A high priority media

monitoring is done across most spreading media. A

monitoring news module utilizes contextual

embeddings to calculate the authenticity of news

articles. Integrating with news outlets, it constantly

fetches articles for review. It utilizes advanced natural

language processing techniques to generate

contextual embeddings for each article that capture

fine-grained semantic information. These

embeddings are then processed through a pre-trained

model that is built on real news articles, allowing the

module to verify authenticity through the calculation

of similarity scores.

Module 1: News Classification — Extracted

news will be classified in this module. Sensitive and

non-sensitive news are classified thus Non sensitive

news gets published. The Classification Module: A

module that employs machine learning algorithms to

classify news articles in various categories, separating

authentic articles from potentially misleading ones

based on the output generated by the module titled

"Enhancing News Authenticity Assessment with

Contextual Embeddings." The classification module

consisted of threshold-based classification to the

Determine authenticity category of article, upon

receiving contextual embeddings and authenticity

scores from the above module.

Articles that score above an authenticity threshold

are classified as authentic, for which the algorithm

has a high certainty, vis-a-vis the reliability and

accuracy of facts in the article. On the other hand,

any article that scores lower than the threshold is

labeled as potentially harmful, requiring additional

examination or confirmation. Figure 1 shows the

Blockchain and NLP-Based Framework for News

Validation and Opinion Synthesis.

Figure 1: Blockchain and NLP-Based Framework for News

Validation and Opinion Synthesis.

Publish Authenticated News: Secure news

authentication. This module sends the Sensitive news

for authentication. After integrity check it will be

supervised learning. News will be published by

authentication. The Publish Authenticated News

module acts as a forum for publishing authenticated

news articles, designed to promote the distribution of

accurate and checked information to the general

population. It works together with the authentication

and classification modules to ensure news articles are

genuine before being published.

When an article is submitted by a journalist or

contributor, the module conducts a full authenticity

assessment using contextual embeddings and

classification algorithms. Articles that surpass this

threshold of authenticity are considered

authenticated and eligible for publication.

The module provides a user-friendly content

management interface where editors can validate,

enrich with context or metadata, and set publish times

for authenticated articles. It also enables collaborative

editing and approval workflows to maintain editorial

integrity. In addition, the module connects with

social media and news aggregation channels,

enabling the dissemination of validated pieces to 【ɑ

】 larger audience. It uses HTTPS to ensure that the

data when published is protected from interception

Synthesizing Human Expert Opinion to Assess News towards News Authenticity

and malicious attacks. If the authenticity score of an

article is greater than a preset threshold, it is

categorized as authentic, thereby establishing a

relatively high certainty that the article is reliable and

factually appropriate. On the other hand, articles with

scores lower than the threshold are designated

potentially misleading and require further inspection

or verification.

4 CONCLUSIONS

A marvel of contemporary media: a delicate

balancing act, ever so close to being toppled by

unforeseen consequences applied nationally. The

proliferation of skewed journalism, motivated by

finance and socio-political agendas, has resulted in

widespread distortion, and added a layer of fog to

news reporting. Issues of socio-political impact,

shrinking freedom of the news terrorism and the

necessity for a basic news evaluation process have

appeared regarding the increasing sulfide of internet

news. There are several factors that encourage fake

news, such as political polarization, motivated

reasoning, and social media algorithms, all of which

play a role in how fake news spreads, and this is not

without consequences. Then the news we read

should come from real-world sources and contain

truthful information. This study heats up the hastst

game of identifying and verifying which news

authenticity in order to tackle the critical challenge of

countering the influence of fake news on public

perception towards general topics and human societal

conversations.

REFERENCES

A. Riaz and F. Zaman, ‘‘Working under the ‘sword of

damocles’: Experiences of journalists in a hybrid

regime,’’ in Masks of Authoritarianism. Singapore:

Springer, 2022, pp. 37–55, doi: 10.1007/978-981-16-

43149_3. management, analysis and future prospects.6

,54(2019). https://doi.org/10.1186/s405370190217.

A. Riaz and F. Zaman, ‘‘Working under the ‘sword of

damocles’: Experiences of journalists in a hybrid

regime,’’ in Masks of Authoritarianism, A. E. Ruud and

M. Hasan, Eds. Singapore: Palgrave Macmillan, doi:

10.1007/978-981-16-4314-9_3.

B. Uddin, N. Reza, M. S. Islam, H. Ahsan, and M. R. Amin,

‘‘Fighting against fake news during pandemic era: Does

providing related news help student internet users to

detect COVID-19 misinformation?’’ in Proc. 13th

ACM Web Sci. Conf., Jun. 2021, pp. 178–186, doi:

10.1145/ 3447535.3462508.

K. A. Ahmed, ‘‘In bangladesh: Direct control of media

trumps fake news,’’ J. Asian Stud., vol. 77, no. 4, pp.

909–922, Nov. 2018.

M. A. Al Mumin, M. H. Seddiqui, M. Z. Iqbal, and M. J.

Islam, ‘‘Neural machine translation for low-resource

EnglishBangla,’’ J. Comput. Sci., vol. 11, no. 11, pp.

1627–1637, 2019.

M. H. Alkawaz and S. A. Khan, ‘‘Use of fake news and

social media by main stream news channels of India,’’

in Proc. 16th IEEE Int. Colloq. Signal

Process.Appl. (CSPA), Feb. 2020, pp. 9397, doi:10.11

09/CSPA48992.2020.9068673.

M. Kowsher, M. J. Uddin, A. Tahabilder, M. R. Amin, M.

F. Shahriar, and M. S. I. Sobuj, ‘‘BanglaLM: Data

mining-based Bangla corpus for language model

research,’’ in Proc. 3rd Int. Conf. Inventive Res.

Comput. Appl. (ICIRCA), Sep. 2021, pp. 1435–1438.

Q. Shishir, ‘‘Reports misleadingly claim Saudi Arabia

plans to remove Islamic oath from national flag,’’ Tech.

Rep., Feb. 2022.

R. Bayes and J. N. Druckman, ‘‘Motivated reasoning and

climate change,’’ Current Opinion Behav. Sci., vol. 42,

pp. 27–35, Dec. 2021.

S. Nath and S. Ray, ‘‘Political campaigning in west

bengal: Violence, professionalisation, and communalis

ation,’’ South Asian Hist. Culture, vol. 13, no. 3,

pp. 305320, Jul. 2022, doi: 10.1080/19472498.2022.20

75208.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES