Real Scientific Research vs. Electronic Search: Why Human Inquiry Still Leads in the Age of Algorithms (Codes and Data)

Real Scientific Research vs. Electronic Search:

Why Human Inquiry Still Leads in the Age of Algorithms (Codes and Data)

Abbas Mokhtarzadeh 

July 8, 2025

“Not everything that counts can be counted, and not everything that can be counted counts.”, Albert Einstein, [1].

A) Introduction

We live in a digital age where information flows faster than ever. With Google, ChatGPT, and a growing ecosystem of AI and machine learning (ML) systems at our fingertips, the landscape of research has transformed dramatically. Instant answers, automated writing, predictive models, and AI-generated visuals now dominate how we access and present knowledge.

It has never been easier to search, summarize, or simulate. But with all this power comes a crucial reality: technology can support research; but it cannot replace it.

For young scholars, students, and professionals across disciplines, it is essential to understand what today’s AI and search tools can do; and what they leave out.

B) What Today’s Tools Offer

Let us begin with their strengths. Modern AI and ML systems, along with advanced browsers, have opened new doors for learning and discovery.

Popular web browsers such as Google Chrome, Mozilla Firefox, Microsoft Edge (now AI-enhanced with Bing and Copilot), Safari, Brave (which blocks ads and includes AI chat features), Arc (designed for focused research), and Opera (featuring its own AI assistant) provide seamless access to the internet’s vast resources.

On the AI front, systems like OpenAI’s ChatGPT offer text generation, brainstorming, analysis, and conversation capabilities; Google’s Gemini integrates web results and live data; Anthropic’s Claude prioritizes safety and contextual understanding; Microsoft’s Copilot is integrated into Office and Edge; and Perplexity AI combines conversational AI with real-time search results., [2,3].

Meanwhile, machine learning platforms such as Google’s TensorFlow, Meta’s PyTorch, Hugging Face’s model hub, and cloud AI environments like Amazon SageMaker and Google Vertex AI provide the infrastructure for developing advanced AI models and research applications, [4,5].

C) What These Tools Can’t Do

Despite their strengths, these tools are not a substitute for genuine academic research. They excel at providing fast answers and accessible summaries; but they do not delve into the deeper layers of credible, peer-reviewed, original scholarship.

They often miss or misrepresent:

1) Journal articles behind paywalls, such as those on JSTOR or ScienceDirect;

2) Primary sources, including historical documents, fieldwork, and lab data;

3) Nuanced scholarly debates;

4) Recent studies not yet indexed;

5) Research published in less-digitized languages or regions;

Here is more details comparison of these unavailable sources to electronic data :

1) Journal Articles Behind Paywalls (e.g., JSTOR, ScienceDirect)

Over 100 million scholarly articles have been published globally across disciplines, [2]. Approximately 30–40% of these articles are open access or freely available through repositories and open journals, [6]. However, about 60–70% remain behind paywalls or require subscription access. Google and AI tools typically index only abstracts or metadata, limiting direct access to full texts.

2) Primary Sources (Historical Documents, Fieldwork, Lab Data)

Archives worldwide hold millions of primary documents and datasets. Only a small portion—roughly 10–20%—has been digitized and made accessible electronically, [7]. The majority, around 80–90%, remain undigitized or restricted to physical archives or specialized databases. These sources are generally unavailable through AI systems or standard web searches.

3) Nuanced Scholarly Debates

There are thousands of ongoing scholarly debates spanning various disciplines, often published in specialized journals or conference proceedings. While some of these discussions are openly accessible, most remain behind paywalls, with less than 40% freely available. AI and search engines may capture summaries or opinions, but deep contextual understanding requires access to full texts.

4) Recent Studies Not Yet Indexed

New studies are published continuously, but indexing in major databases can take weeks or months. Many recent papers are shared initially on preprint servers such as arXiv or bioRxiv, which are open access but have not undergone peer review. Peer-reviewed versions often remain paywalled. AI models trained on data before a cutoff date do not include the latest research, and search engines may lag in indexing these new publications.

5) Research Published in Less-Digitized Languages or Regions

A significant portion of global research is published in languages other than English, including Chinese, Spanish, Russian, Arabic, and others. Digitization and indexing rates for such research vary widely, with many regional journals and theses poorly digitized or not indexed at all. Consequently, AI systems and search engines have limited coverage of these materials, especially in non-English languages or smaller academic communities.

The original context, which is often lost in compressed summaries.

Moreover, AI systems face significant challenges in critical interpretation and discerning reliability, often falling prey to circular references, misinformation, or disinformation that frequently circulate within data sources and databases. These issues undermine the trustworthiness of the outputs generated by such systems.

Additionally, AI struggles to effectively evaluate the credibility of sources and to navigate conflicting or contradictory information; capabilities that are fundamental to the integrity and rigor of thorough, high-quality research. Unlike human researchers, who apply contextual judgment, skepticism, and nuanced understanding, AI remains limited in its ability to question, cross-verify, and synthesize complex or ambiguous data, making genuine scholarly inquiry indispensable in the age of machine learning and automated data analysis.

D) The Risk of Overreliance

Overreliance on AI and search engines can lead to shallow understanding, inaccurate citations, outdated data, overconfidence in machine-generated conclusions, and ethical issues like plagiarism or bias.

As a researcher, your role is not simply to find answers but to question, validate, and analyze; not to accept information at face value.

E) What Real Research Looks Like

Research is not just about finding answers; it is about asking better questions. It requires accessing scholarly databases such as JSTOR, PubMed, and Scopus; reading peer-reviewed papers; reviewing primary sources and original datasets; engaging with opposing viewpoints; and applying critical thinking and methodological rigor, [5,6]

AI tools assist in preparation but cannot replace these essential steps.

F) How to Use Technology Wisely

I) Use AI/ML -----> then, Follow Up With Brainstorming research questions Reading published studies on the topic

II) Summarizing long texts -----> then, Verifying summaries with original sources

III) Outlining essays or presentations -----> then, Supporting ideas with peer-reviewed evidence

IV) Exploring unfamiliar topics -----> then, Consulting expert-authored books and articles

V) Speeding up repetitive analysis -----> then, Reviewing data and interpreting results

and many more.

G) Your Role in the AI Era

You are part of a generation with access to tools past scholars could only imagine. However, tools alone do not make good research; your curiosity, rigor, and ethics do.

“Real research begins where the browser ends.”

Embrace AI to explore, test, and reflect; but never stop there. Be the one who digs deeper, asks more, and brings human insight to a machine-driven world.

References (selective)

1) Albert Einstein, quoted in William Bruce Cameron, Informal Sociology: A Casual Introduction to Sociological Thinking (New York: Random House, 1963).

2) Peter Suber, Open Access (Cambridge, MA: MIT Press, 2012), A concise introduction to the basics of open access, describing what it is (and isn't) and showing that it is easy, fast, inexpensive, legal, and beneficial https://mitpress.mit.edu/books/open-access.

3) Aileen Fyfe et al., Untangling Academic Publishing: A History of the Relationship Between Commercial Interests, Academic Prestige and the Circulation of Research (University of St Andrews, 2017).

4) Michael Goodchild, “The Quality of Big (Geo)data,” Dialogues in Human Geography 3, no. 3 (2013): 280–284, https://doi.org/10.1177/2043820613513392.

5) Emily M. Bender and Alexander Koller, “Climbing Towards NLU: On Meaning, Form, and Understanding in the Age of Data,” Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (2020): 5185–5198, https://doi.org/10.18653/v1/2020.acl-main.463.

6) Heather Piwowar et al., “The State of OA: A Large-Scale Analysis of the Prevalence and Impact of Open Access Articles,” PeerJ 6 (2018): e4375, https://doi.org/10.7717/peerj.4375; Dimensions database, https://www.dimensions.ai/.

7) UNESCO, Archival Digitization: Global Trends and Challenges, UNESCO Report (2020), https://unesdoc.unesco.org/ark:/48223/pf0000372353

& following reference that is not cited in article above as an example of copyright and Plagiarism in our digitalised era for further study and reference:

 Dryden, Jean. “The Role of Copyright in Selection for Digitization.” The American Archivist 77, no. 1 (Spring/Summer 2014): 64–95. doi:10.17723/aarc.77.1.3161547p1678423w.