Abstract

This paper proposes to exploit ChatGPT and its counterparts in a smart, safe, and swift manner, in order to boost human intelligence. Drawing on my expertise in the field of information, I  examine the informational and cognitive skills required to make the best use of these tools and, in general, of any information. Among these:

  1. The importance of cultivating a critical and skeptical mindset when working with any information.
  2. The capacity to effectively process a variety of channels and viewpoints to counterbalance any shortcomings in sourcing and biases.
  3. The ability to scan, fact-check, and quickly identify the intended purpose of any piece of information.
  4. The art of transforming information into knowledge, by contextualizing and connecting any in a meaningful manner.

Developing these competencies is one way to maximize the benefits of these tools, minimize their impacts, and enhance our HI, or Human Intelligence. Developing these skills could benefit everyone, particularly our educational system. Rather than focusing on content (most school programs are still structured according to subject matter, such as history, biology, chemistry, English, etc.), concentrating on the competencies necessary to handle any is an alternative that may restore a sense of meaning to education. To avoid hearing « Well, why should I go to school, everything is available on the internet anyway».

Introduction

Millions of people google and (almost always) find information. With “the tsunami of cheap AI content” invading the internet (to quote The Verge, July 2nd, 2019), what about quality information, which is so essential to making well-informed decisions?

In the era of ChatGPT (-3, in November 2022; -4, in March 2023), Google Bard AI, Bing AI Chat, ChatSonic, Claude, Ernie Bot, and many other Large Language Models, finding and utilizing information has become more challenging than creating it, which these assistants do plentifully and easily. It has become increasingly difficult to navigate through all the data that these tools (and other electronic ones) generate, as James Vincent pointed out in  How AI is killing the old web and the new web struggles to be born (The Verge, June 26th, 2023). 

With forty years of experience in the research and information fields, through different sectors and roles (academic and professional), with a background in semantics and cognitive sciences, I am the author of the “Knowledge is Power” series (available on most Amazon marketplaces). In the second volume, I focus on the competencies necessary today to be an efficient information re/searcher.

This present paper also concentrates on informational competencies. I examine how to use ChatGPT (and its like) and whether these tools can be useful knowledge assistants for enhancing our human intelligence. Let us start with some of their drawbacks. 

What is Wrong with ChatGPT (and its like)?

ChatGPT (and its like) can best be appreciated by testing them out on a subject in which you have some expertise. This is what I did, asking ChatGPT about the competencies necessary to be an efficient information researcher, which is my field of expertise.  A well-structured and clearly expressed presentation of the issue was returned, with some general points, nothing very original, but a valuable overview to begin with. Nevertheless, none of the elements were spontaneously documented, and when questioned about its sources, the bot remained vague and did not provide any specifics, as is the norm in mainstream journalism, research, and debate.

For ages, publishers have placed a high value on sourcing. And in many parts of the world, students are painfully reminded to carefully back up any of their writings and not to plagiarize. Many systems have been developed to catch the ones who do it. But what about ChatGPT? Isn’t it a plagiarist, since it does not acknowledge its sources? The press and the web have often discussed the issue of copyright for the information these bots use without attribution. Early July 2023, ChatGPT faced its first Copyright lawsuit, according to the Guardian. Two authors attacked the bot for “unlawfully ingesting their books”. Already, on Dec. 21st, 2022, a Forbes article wondered who ultimately owns content generated by ChatGPT. Haldanes, a Hong Kong law firm wrote (on June 13th, 2023) along the same line “Feed the robot; Starve the copyright owners”. In addition to discussing the legal aspects of AI generative models, the latter emphasized the fact that (so far) these models depend on the creativity of human authors. Taking this issue at a broader level, one might wonder if innovative and creative content may be at risk. What type of original information will be available in a couple of years, if these AI tools generate malinformation, disinformation, and misinformation (to use Claire Wardle’s various types of information disorders) and with little to no possibility to fact-check anything?

Given the extensive usage of these bots, does this mean that the fundamental principles of properly documented information and knowledge are obsolete? Are we at a turning point? This is an important question to ask ourselves. Furthermore, ChatGPT and its like exacerbate the problem of fake news, false information, and data pollution on the Internet, problems that have become increasingly prevalent in recent years.

In spite of providing general introductions to a wide range of topics, and based on our example, fact-checking what it said was challenging to impossible. And when asked, the sources it quoted, besides staying generic, were essentially American. According to OpenAI’s blog, ChatGPT was trained on the collective writing of humans throughout the world up until the end of 2021. The  “collective writing of humans” is vague.  What does this include and does not include? What about the value and quality of the information it generates? James Vincent (above quoted) submitted a ChatGPT text about trekking shoes to a (human) expert who said the bot’s answer was essentially filler “there are a bunch of words, but no real value in what’s written … Taking advice from AI will most likely result in hurt feet on the trail.”

Furthermore, what if ChatGPT answers were hallucinations, to use a concept popularized by Google AI researchers in 2018, about large language models that hallucinate? In a March 13th, 2023 paper from the IEEE (Institute of Electrical and Electronics Engineers), Craig S. Smith discussed how Hallucinations Could Blunt ChatGPT’s Success > OpenAi says the problem’s solvable. Yann LeCun says we’ll see. Moreover, much of the material on which ChatGPT relies is the product of humans, and as the adage says errare human est, i.e. to err is human. Does ChatGPT have some solid mechanisms to screen the numerous wrongs that invade the web, which Google has often complained about? Also, what if these generative tools have corrupt or unrighteous goals and ethics? In one word, can I reasonably trust what these systems are telling me?

OpenAI (i.e. ChatGPT’s developers) discuss their mission, values, and principles in their About webpage (as of July 6th, 2023): “to ensure that artificial general intelligence benefits all of humanity”. A decent objective that stays nevertheless vague; what about their true intentions, their ethical considerations behind “benefiting all of humanity”? Which humanity and culture/s? They research “how to align their generative models with human values”, but what are the human values on which they base their work? Is any control in place for any unrighteous use of their results to generate disinformation? It is a concern that many share. According to data from Google Trends, searches for “is ChatGPT safe?” have increased by a massive 614% since March 16th, 2023 (quoted by the International Accounting Bulletin, March 31st,  2023). Furthermore, in our time of Diversity, Equity & Inclusion (DEI), what about multiple points of view? Having no easy access to OpenAI’s values and how ChatGPT generates its text without backing it up is concerning. It gives OpenAI enormous power.

ChatGPT : A Knowledge Assistant among Many

By contrast, Wikipedia is a collective endeavor. Even if it is often disputed by the scientific community (nevertheless a “science shaper” according to a 2018 MIT Sloan Research paper, from Neil Thompson and Douglas Hanley), a number of its posts are written by a variety of contributors, and this provides some safeguards mitigating biases. In addition, often, the history of how the posts are created is openly available under the tabs: Talk and View History. Check for example Wikipedia’s Talk tab of the “ChatGPT” entry; it starts (as of early July 2023) with an “Edit war on ideological bias of ChatGPT” which presents many controversial positions, duly quoted. Wikipedia documents the generation of much of its content and provides a list of sources at the end of many of them, allowing any researchers to fact-check the posts and investigate further (this often represents a great source to get started).

In regards to Google, it has taken steps in recent years toward slightly more transparency, providing insights into its results. Click on the three dots, next to the first line of each hit. You can see an About this result window, with details on your search and its link with the result. Launched mid-2021, this is still in the Beta version as of July 2023. It may not be perfect, but it is a small sign of improved sourcing and greater transparency, especially when it comes to personalized results. Finally, contrary to ChatGPT, Google offers dozens of results, not just one. While sometimes they may not be very different from one another, the fact that there are many of them engages us with a variety of perspectives.

Not being a native English speaker, I am an eager user of AI personal writing assistants (such as Wordtune for example). And I appreciate them suggesting various options to improve my wording. I can choose the one that corresponds the closest to what I want to say, besides improving my English.

Along the lines of increased transparency, we can also mention the EU GDPR laws (General Data Protection Regulation), the Digital Services Act, and the Cookie Laws, which require the tech giants and any website owners to provide transparency regarding their backstage procedures. This is why, since 2022, internet users are frequently asked if they consent to have their personal information tracked and used.

ChatGPT’s limited results and lack of transparency appear to be contrary to the current trends. It may be perceived as a step backward. Furthermore, the fact that it formulates its answers so elegantly, authoritatively, and humanlike intimidates many and does not motivate any to challenge and further explore. It is similar to what happens when we speak with medical professionals. It requires strong minds, confidence, and/or expertise in the topic to question ChatGPT (and doctors). Rather than expanding knowledge, the bot narrows it, cultivating little critical thinking. A monopoly of knowledge is no good news for society, especially if it falls into the wrong hands. OpenAI (and the like) carry a great deal of responsibility and power towards society.

A Selection of Noteworthy Opinions on ChatGPT

Mid-May 2023, Sam Altman, OpenAI’s CEO,  called for US lawmakers to regulate artificial intelligence. Quoting a May 17th, 2023 BBC article, Altman argued that these tools can create incredibly human-like answers to questions – but can also be wildly inaccurate. “I think if this technology goes wrong, it can go quite wrong…we want to be vocal about that,” Mr Altman said. “We want to work with the government to prevent that from happening.” Even though he certainly speaks for his parish (ChatGPT is ahead of the game, so he might want to regulate the process to maintain his leading position), we should be worried.  The Time published on June 21st, 2023, an in-depth article on him and the opportunities and threats of AI.

According to Cade Metz (New York Times on May 1st, 2023), The Godfather of AI Leaves Google and Warns of Danger Ahead to be able to freely speak out about the risks of A.I. This godfather is Geoffrey Hinton, an artificial intelligence and deep learning cognitive-psychologist and computer scientist pioneer. He considers that ChatGPT’s intelligence is absolutely inhuman and that it is hard to see “how you can prevent the bad actors using it for bad things”. In another article, published on June 10th, 2023, Metz expands on AI destroying humanity. And Eliezer Yudkowsky recommended, in a strongly worded Time article (March 29th, 2023), to shut it all down; pausing AI development is not enough. “If somebody builds a too-powerful AI, under present conditions, I expect that every single member of the human species and all biological life on Earth dies shortly thereafter.” Already early 2000s, he had raised concerns about AI destroying humanity.

End of March 2023, a group of AI experts including Elon Musk and many well-known figures voiced some strong concerns regarding AI and suggested a moratorium. By the end of August 2023, more than 33’700 had signed this Pause Giant AI Experiments open letter.This letter “… urged A.I. labs to pause work on their most advanced systems, warning that they present profound risks to society and humanity”. Nevertheless, the same controversial Elon Musk announced on July 12th, 2023, launching his new AI company, xAI, kicking off the race against OpenAI, which he had cofounded in 2015, but left in 2018. Did he need this moratorium to launch his AI company?

Linked to this Open Letter, the same New York Times article from March 29th, 2023 described how these bots hallucinate, and what is wrong with them. They “live in a world of language,” said Melanie Mitchell, an A.I. researcher at the Santa Fe Institute. That world gives them some clues about what is true and what is not true, but the language they learn from is not grounded in reality. They do not necessarily know if what they are generating is true or false.”

Besides this famous Open Letter, the brief “Statement on AI Risks” (from the Center for AI Safety, May 30th, 2023) incorporated the notion of human extinction straightforwardly. “ Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” It has collected signatures from thousands of prominent figures, such as Geoffrey Hinton, Sam Altman, and Bill Gates.

Another solid signal came from the Italian government, which surprised many by banning ChatGPT over privacy concerns in early April 2023, even if it lifted this ban at the end of the same month. Other countries are banning ChatGPT or discussing it (websearch chatgpt ban countries, for an updated list). The European Parliament has been struggling with its AI regulation for more than two years (according to The Verge, June 29th, 2023). Finally, in the US, in mid-July 2023, the Federal Trade Commission (FTC) opened an investigation into OpenAI about ChatGPT possibly harming consumers through its collection of data, and more generally its security practices.

This is just a very small selection of measures and viewpoints on ChatGPT and AI. Since ChatGPT’s official launch on November 30th, 2022, it has been generating thousands of clear-cut opinions, from extreme fear (of human extinction) to fascination (the ultimate state of intelligence), and from a wide range of figures all over the world. These AI knowledge assistants do not leave anyone indifferent, and many companies are frantically competing for access to this lucrative market. It has been a long time since we have seen something this controversial in the field of technology. This is healthy. It is a pity that ChatGPT does not trigger the same level of critical thinking in its users, as its debate does. Finally, we are seeing early summer 2023, a shift in what is published about these bots. Several practical articles are now being published that describe how to use ChatGPT for different purposes, whereas they were more theoretical in the early part of 2023, focusing on the suitability and value of the tool. Does this indicate that the acceptance rate is increasing?

How to Make the Most of ChatGPT (and its like)

No one can ignore ChatGPT and its like. The pace at which ChatGPT-3 has been onboarded is impressive.  In only five days, it reached 1 million users; by comparison, it took 75 days for Instagram and 10 months for Facebook, to do this, according to tooltester.com. And the billions of dollars they have received so far, indicate these bots are here to stay. Microsoft besides investing billions into ChatGPT has already included parts of it in its products. Google devised  “Radical Search Changes to Beat Back A.I. Rivals” according to the New York Times (April 16th, 2023).

Related to ChatGPT’s poor to inexistent backup and sourcing, and considering how it is constructed, it is difficult to imagine that this fundamental aspect will be significantly altered. Consequently, we have to do with these AI assistants. Let us see how to take advantage of them safely and wisely, for humans to enhance their intelligence, make better decisions, and save the planet.

With the abundance of information available today, the ability to quickly scan, question, and assess any information becomes increasingly important. Ask yourself: What are the author’s intentions, Why am I considering/adding this new piece of information, What does it bring to the rest, How does it connect with the topic and the other elements, finally, Is it relevant and reliable? Quality re/search can (almost) never be achieved through a single ChatGPT or Wikipedia contribution, but rather through the combination of many sources. It is essential to check with others, even if it is only for confirmation purposes. Contrasting or divergent views can also be incorporated since they tend to enrich the argument and the exchange of ideas.

Like in investment, diversifying your sources, tools, and channels is key; it is a way to mitigate the risk of using false information. Try to use at least three sources/channels, and more if they diverge. Greg Brockman, one of Open-AI’s co-founders, suggests using ChatGPT as a brainstorming partner (Ted Talk (April 20th, 2023). This usage also adheres to the concept of considering multiple perspectives. Much like with Wikipedia (but without its rich references), ChatGPT may be used to get started i.e., to gain a concise overview of the issues at stake, without forgetting, however, to move to other viewpoints and sources and to stay highly critical. The aim is to confront different pieces of information, make their intended purpose explicit, and connect them. In addition to making sense and acquiring meaning, they become easier to work with.

If you cannot find any satisfactory answers in your research, no problem, explain why and mention the sources you used. An absence of results can be as informative as the presence of many, to those who know how to interpret it. Do not try to twist your results, but contextualize and justify what you deliver. By giving your readers the context of your argument/research, you allow them to accept or reject your results with sufficient evidence, and most important, to build from them. And in fine, to make decisions properly supported. Sourcing, contextualizing, and giving a sense of purpose to the various pieces of information that comprise your findings is critical to its delivery and use. Unfortunately, ChatGPT with its well-written answers does not encourage users to apply and develop these skills.

In many parts of the world, the state of mind toward information, which can be traced back to Gutenberg and the invention of printing, needs to change. It is not because it is printed that it is true and set in stone. There is a need for a paradigm shift. Multimedia and the Internet have started to shake this belief, by offering a multitude of information channels, without however giving the tools (and the training) to handle this abundance of right and much wrong information. In front of this superabundance, many of us are overloaded, lost, and often paralyzed. Being given just a single answer is so much more comfortable. This is probably why « influencers » have been so popular lately; indeed, they tell us what to wear, eat, consume, think, and be. Without comparing ChatGPT to influencers (at least content-wise), we can see some similarities in the role: providing single and authoritative answers to almost everything. They comfortably fill our information gaps, and reduce our uncertainty and stress. These little bots, which speak to us so nicely, make us feel good, much like dopamine. Why should we put their contributions into question?

ChatGPT is attractive because it is convenient. We have never been given access to so much information and knowledge, so quickly, so nicely and authoritatively presented, and so human-like. Just as convenience food is ready for consumption, ChatGPT’s single answers require little effort to be consumed. By contrast, a diversity of information is demanding; it needs some multi-processing skills with a critical mind. This is probably the greatest danger of ChatGPT and its like: not promoting and nurturing a critical mind and the capacity to process, analyze, synthesize, categorize, organize, question, assess, establish priorities, and scan many pieces of information.

Older generations, of which ChatGPT conceptors are part, have gone through this exercise and are capable of navigating through large volumes of information. But what about the generations to come? They are particularly at risk since they do not perceive any obvious need to develop these skills. Without the need, they are unmotivated to access a variety of information. ChatGPT will have the answer to all, so why even go to school? It is not so much a question of content (there is plenty of it all over), but rather of the skills that are required to efficiently process and utilize it. The educational system continues to be heavily structured by subject matters (history, geography, English, chemistry, etc.). And this is in some ways out of sync with today’s world. New generations would be better prepared for their digital world if they were taught how to navigate, process, analyze, prioritize, and question their information and knowledge universe. It would also build on a concrete need, so much more powerful for learning.

As a rule, aim at Rich information, one of the R of « 4R  quality information », the others being Recent, Relevant, and Reliable. Nevertheless, accept not to be exhaustive.  Considering the vast amount of information available on (almost) any topic, to guarantee a thorough review of all available information is unrealistic. By documenting, contextualizing, and explaining your selection of information, you will be held accountable for what you have done, and your readers will be able to accurately position your findings. Setting a timeframe when you search, can also help to stay focused to avoid going overboard on the endless internet avenues. Use good sense and keep in mind Vilfredo Pareto’s 80/20 principle, with the first 20% of most projects bringing most results, and the last 20% being highly resource-intensive, with often little extra value. Websearch “pareto principle” to know more about how this applies to the management of many projects. 

Accepting that the plurality of information is healthy, that any can (and should) be challenged, is a significant shift in beliefs. And in our world where speed and gaming dominate, taking time to visit a variety of information might not be very popular. However, it is imperative to embed critical thinking into the way we approach information, at all levels, and teach it widely and early at schools. The first message is straightforward to grasp: relying on a diversity of tools, channels, sources, and results, even if it is more time-consuming. In our world of misinformation and fake news, it is the only way to mitigate the risks of using bad information. The second message is more challenging: juggling all of them, adopting a critical and skeptical attitude to each, and not taking any for granted… even this paper.

Any piece of information, a ChatGPT answer, a Google hit, or any prestigious paper needs to be questioned, even if it is for a few seconds. It is worth recalling that even two eminent journals, The Lancet and The New England Journal of Medicine, published and then retrieved poorly supported papers during the early days of COVID-19 (mid-2020). It concerns all, but especially scientists, researchers, journalists, students, educators, analysts, politicians, and librarians. We should cultivate a culture of skepticism by constantly questioning information. This may be a substantial shift in attitude, but it is necessary if we are to make the right decisions.

Even if disagreeing is not always accepted in some cultures and companies, people must be allowed and encouraged to do it, without others taking it personally (and without being imprisoned or fired). Turning disagreement and the exchange of ideas into a play is a way to minimize the personal attack side of things, provided the exercise is conducted honestly and constructively. Developing critical thinking skills (teaching and applying them) will be our means of remaining as intelligent as, possibly more intelligent than these AI-generated systems. And to restore HI, Human Intelligence. In other terms, to stay free thinkers.

OpenAI (in its About webpage) mentions that its systems are « generally smarter than humans ». Even if « I do not consider myself to be omniscient or all-knowing » (said ChatGPT when asked, on April 19th, 2023, if it knew everything), it vehicles an image of omniscience. The mythology is full of gods and goddesses of knowledge: Athena (Greek), Saraswati (Hinduist), Odin (Norse), the Ibis bird (Egyptian), etc. Omni-science fascinates but also frightens, as it conjures up images of mind dictatorship, controlled thought, and “single thinking” (“la pensée unique” in French). Yes to use these tools as starting points, as one source among many others to build from, but no, to use them as the Ultimate word. Even if they are very convenient. Our world desperately needs a plurality of documented views. Google despite its zero-click strategy, and even if it has many biases (remember the filter bubble denounced by Eli Pariser, the internet activist, in a May 2th, 2011 Ted Talk, which comforts us in our thinking, isolating us intellectually), at least offers its users various results (about 200, in July 2023; in 2019, it was about 300) to choose from, many of them with sourcing details. The only way to mitigate the risks and biases posed by these bots is to build the ability to handle a diversity of information.

Conclusions

I turned to ChatGPT (on June 12th, 2023) to have its suggestions to « be a free thinker ». Its answer was detailed and well formulated (as usual). Try it for yourself. I shall only quote its last point « Remember, becoming a free thinker is an ongoing process that requires continuous learning, self-reflection, and an open mind. It’s about embracing intellectual independence while respecting and considering alternative perspectives.” We can only agree with this, even if ChatGPT does not foster free thinking skills.  

I then asked if it believes it develops free thinkers: “ As an AI language model, my purpose is to provide information, assist with tasks, and engage in conversations. […] However, it’s important to note that the responsibility for cultivating free thinking ultimately rests with the individuals themselves. It requires active engagement, exploration of diverse sources of information, and a willingness to question and evaluate ideas independently. Interacting with an AI like me can be one of the many resources individuals use in their journey towards becoming free thinkers, but it is not the sole factor in their development.” A fair and detailed answer, even if it does not see any responsibility for the fact that it does not encourage its users to go to “diverse sources of information … many resources”. Unfortunately, it advocates the opposite of what it induces.

Will the convenience of ChatGPT single answers outweigh “quality” information and knowledge? Will these new AI knowledge assistants become the new “Google” used by half the world population?  Will they benefit all of humanity, i.e., will they basically make us better and happier human beings? Will they help us in our “pursuit of happiness”, which is one of the principles enshrined in the US Declaration of Independence and other founding texts?

Even if these tools have made their way into our lives in less than a year, we can still have the last word. The large number of opinionated articles and posts published virtually daily, the moratorium, the Statement of AI Risks, and the countries that have banned it, are encouraging us to rethink what we want: true knowledge assistants that help us to be better and to grow, or dictators that impose their truths? By early summer 2023, a number of articles indicated that the adoption of ChatGPT was slowing. Quoted by CNCB “Bank of America analysts estimate visits to ChatGPT were down about 11% on a monthly basis to just over 51 million visitors per week, or only about 2% of Google’s estimated web traffic.” Is it because the novelty wears off or because schools are on holiday? End of August 2023, some posts raised concerns about the financial stability of OpenAI. Regarding the decline in users,  tooltester.com provides many interesting statistics on ChatGPT.

Finally, when it comes to experience, common sense, emotional intelligence, and consciousness, humans are still ahead of the game, as ChatGPT reminds it when asked about its omniscience on April 19th, 2023:  “I do not possess consciousness or the ability to fully comprehend human experience in the same way that a human being can”. Consciousness is difficult to model, but central to information processing. Quoting an interesting article on the theories of consciousness (from the New York Times, July 1st, 2023), “consciousness is a byproduct of the way we process information… consciousness is the global availability of information”.

Consciousness is what makes humans unique. This is our opportunity to keep our world human and humane. Rather than lament and weep, let us be serious and learn to master our information universe and its related tools.  

The informational competencies that I highlight in my three-handbook series “Knowledge is Power” are in the background of what I discuss in this present paper. Finding and utilizing quality information, wherever it comes from, requires staying focused, being intellectually agile, and exercising a critical and skeptical mindset, while being able to process a variety of information (including opposing ones). The second handbook “Information Research Techniques” presents a large array of techniques to search, assess, and use any information. For more, refer to any Amazon marketplaces, under christine gardiol or to my website, and do not hesitate to leave me your comment.

 

Annex 1 : Informational competencies, according to ChatGPT

« What are the competencies necessary to be an efficient information researcher » is the question I submitted to ChatGPT on March 23, 2023. Here is its answer.

To be an efficient information researcher, you need a combination of skills, knowledge, and attitudes. Here are some of the competencies that are essential for this role:

  1. Information literacy: This includes the ability to identify information needs, locate and evaluate sources of information, and use information ethically and effectively.
  2. Research skills: You should be able to develop research questions, design and implement research projects, collect and analyze data, and communicate findings effectively.
  3. Critical thinking: You should be able to analyze and evaluate information critically, and to recognize biases, assumptions, and logical fallacies.
  4. Technical skills: You should be proficient in using digital tools and platforms for information retrieval, organization, and analysis, as well as in using relevant software programs.
  5. Communication skills: You should be able to communicate effectively with a range of stakeholders, including clients, colleagues, and the public, using a variety of media.
  6. Project management skills: You should be able to plan and manage research projects effectively, including setting priorities, meeting deadlines, and coordinating resources.
  7. Flexibility and adaptability: You should be able to work effectively in a variety of settings and adapt to changing circumstances and technologies.
  8. Persistence and attention to detail: You should be able to persevere in the face of challenges and pay close attention to detail in order to ensure accuracy and completeness of research findings.

Overall, an efficient information researcher should be able to think critically, conduct research effectively, and communicate findings clearly and accurately. They should also be able to adapt to new technologies and be willing to learn continuously to stay up-to-date with the latest information research methods and tools.

I then asked ChatGPT to quote its sources.

As an AI language model, I don’t have sources in the traditional sense, as I have been trained on a vast corpus of text from the internet. However, I can provide you with some examples of sources that discuss the competencies necessary for efficient information research.

  1. The Association of College and Research Libraries’ Framework for Information Literacy for Higher Education: This document outlines the information literacy competencies that students should acquire in order to be successful in college and beyond.
  2. The American Library Association’s Competencies for Information Professionals: This document identifies the core competencies that information professionals should possess, including information literacy, research skills, and communication skills.
  3. The Society of Competitive Intelligence Professionals’ Core Competencies for Competitive Intelligence Professionals: This document identifies the competencies necessary for professionals working in competitive intelligence, including research skills, critical thinking, and communication skills.
  4. The International Association of Scientific and Technological University Libraries’ Guidelines for Information Literacy in Higher Education: This document provides guidelines for developing information literacy programs in higher education, including the competencies that should be emphasized.
  5. The Special Libraries Association’s Competencies for Information Professionals: This document identifies the competencies necessary for information professionals working in a variety of settings, including research skills, critical thinking, and communication skills.

I hope this helps!

Annex 2 : 13 competencies necessary in 2023 to be an efficient researcher – an essential guide

The following list of competencies is based on hundreds of research requests conducted, over the last 25 years, that I have done at a competitive intelligence office and an international financial company. Researching a wide range of issues and using a very large set of information tools, besides deep digging into the general web. 

No. 1: Needs elicitation, i.e. going beyond the mere words of a research or request for information, to uncover the underlying needs. Investigating an issue below the surface allows for precise identification of what needs to be found, and then for clear directions on how to proceed.

When you seek information for others, a valuable technique for achieving this needs elicitation is reformulation.  If I understand you correctly… is an excellent way to begin reformulating. You may want to ask them about their “working hypothesis”, the results they anticipate getting in an ideal scenario, and how they intend to use your findings. As a result of these indirect questions, they are usually encouraged to open up and reveal more about their true needs.

When conducting your own research, keep an audience in mind. By doing so, you gain a sense of direction and can select resources and information more effectively. Break your research project down into smaller units, and then analyze each one separately. You may find it easier to formulate your needs in the form of questions. You may also just express them as my audience needs to know about this or that…, or I need to find the answer to this and that…, or in an ideal world, I would like to find this and that. This will force you to dig deep and be concrete and action-oriented.

Take time to do this, it pays off in the long run. Resist googling right away. Acting upstream, i.e., carefully defining and setting your research needs and goals will make it easier downstream, i.e., when searching, analyzing, and putting together your research results. Your research process is streamlined and many potential pitfalls are avoided in this way.

No.2 : When you reformulate a request, you also need active listening competencies, to “read (listen) in between the lines”. In a nutshell, it means carefully listening to the requestors’ message, their needs, intentions, assumptions, and expectations. Ask your requestors any questions at this early stage, and until it is as clear as possible since they are usually more inclined to give details at this point rather than later on in the research process. Nevertheless, ensure that your strategy is tailored to the time availability of your requestors. Use good sense.

No. 3: It is also important to be able to quickly identify some best resources, information systems, channels, authors, and sources, that you will use to find your answer/s. A way to do this is to ask yourself, which authors/systems/entities might most likely have spoken/written/published on what you need to find. Alternatively, you may screen resources and channels from the perspective of your target audience. Although Wikipedia posts are of uneven quality, they are worth reviewing for the often long list of sources they give at the end. Resource-wise, all of these are excellent starting points.

No. 4 : To determine the most effective entry points for searching, i.e. the concepts and keywords that best represent the issues that need to be addressed is the next competency you need to have. Try to think from specific to general and vice versa, or to use an image, see the trees and the forest, and move between the two perspectives. 

The idea is to develop a search strategy and plan. It is so easy to get lost in the information galaxy, with the growing number of poor-quality web pages generated by AI. The keyword here is: prioritizing. Basically, it is arranging each element of the request according to its importance, impact, or urgency. Each belongs to some smaller sets which depend on each other. As a result, the whole becomes more meaningful. In contrast, when the information is flat and disconnected, our brain has difficulty processing and retaining any, as well as transforming it into knowledge. Don’t we say that memory is all about connections? For learning to take place, neurons in the brain must establish connections between themselves, with synapses serving as the key agents in this process.

Establishing connections by prioritizing information thus helps our brain to function. Concretely, when you start your research, you should interrogate and decompose what you need to find. Extract the different issues/sub-issues, and sort them into three or more groups: priority, secondary, and tertiary. Then identify what connects each of them. Establish your list of the different issues to be searched, and link them with the most promising information resources, search terms, and search strings. By doing so, you will have some solid foundations and directions to get you started.

No. 5 : Keyword selection. As we have all experienced, finding “good” information is greatly influenced by the choice of keywords; a word rather than another may generate very different results. Words can have various meanings (polysemy) and multiple connotations (usages) and it is essential to remember this when searching for information. Move from specific to general  (trees and the forest), and the reverse. Think about families of words, synonyms, antonyms, and tagging of valuable resources. Check extensively thesaurus (many exist online), and any of the keyword tools and technologies used by the SEO (Search Engine Optimization) sector to build websites. They are as concerned with keywords as you are; you are at each end of the spectrum.

Dare to go beyond the terms of the request. A successful choice of keywords is also heavily dependent on how carefully you identified the needs of your research (back to competency no. 1). If these are unclear or fuzzy, you will receive irrelevant and unreliable information. Also, never stick to Google only, diversify your information tools and resources. Make sure to critically examine the results and be prepared to question the keywords you selected. Finally, try to keep track of the most important search strings you use (i.e. the combination of keywords, functions, and commands). By not doing this, you may find yourself going in circles. A successful keyword search requires agility and rigor. 

No. 6 : Timeframe. Setting a timeframe to avoid going overboard is also essential, especially in a professional setting. The time (and resources) invested for a search have to be related to the value of the information that is needed; spending 10 hours for « nice to have » information might not always be justified.

Think about the return on investment of the search, and Vilfredo Pareto’s 80%/20% rule (the last 20% being very labor intensive, and offering a low return on investment). Determine as well if the request will need to be updated later (in this case, document it precisely to be able to rerun it easily later).

No. 7: Being able to “translate” key concepts and key words, into the search language of the different information systems you plan to use (i.e. basically into effective “search strings”) is critical. Besides keywords, most search engines use commands, functions, wildcards, and syntactic rules to specify a search.

AND, OR, NOT (i.e. the Booleans) are used by most information retrieval systems. AND is usually the default one (for example with Google); this means, you do not need to type it; the system understands it is AND. Aside from Booleans, there are several other commands and functions to build search strings that can greatly influence the results you receive. For example, some systems offer:

  • NEAR (plus a number). It links 2 keywords (or phrases) and specifies how far apart they should be.
  • WC for word count. It specifies the length of the document.
  • ATLEAST (plus a number). A keyword needs to be in the document a certain number of times.
  • HD. My keyword/s need/s to be in the headline.

Using the above-mentioned commands, the assumption is that if one or more keywords are close to each other, and in the headline paragraph, and used many times in the article/post, and with the latter being over 1000 words long, the hits will be highly targeted. If you can also determine the timeframe, the geographical scope, and a group of sources (for example, all the sources of a country, or all major business press, etc), the search, and consequently the results will be even more focused.

While today there is more standardization of search commands and functions than two decades ago (largely due to Google’s influence), there may still be some subtle differences that may produce wildly different results. What makes search engines unique and powerful is their ability to fine-tune search strings (through commands and functions). If you want to achieve highly focused results, it is important to know which ones are used by your regular information systems.

No. 8: Keeping Google’s biases in mind (its zero-click strategy and efforts to personalize its hits, which skew results), it nevertheless provides a user-friendly template for conducting targeted searches with ease. It is the Advanced Search screen which can be found in the Settings menu (bottom right of most Google homepages). Any heavy Google users, interested in searching more precisely but not in learning its search functions, should become accustomed to using it. 

Advanced search screens are very handy, as they allow anyone to build precise searches without having to remember complex commands and syntax.  They are much more frequent with professional information systems. Try spotting them when onboarding any new information systems. Even basic email systems (such as Yahoo) offer it. Twitter has one as well. Sometimes it is a bit hidden, but be tenacious and look for it. To know more about it, websearch under “advanced search” with or without the name of the search engine or the information system. Check as well as YouTube tutorials.

Sadly, not all search engines offer such a search template (it is sometimes also referred to as Power Search). It is increasingly being replaced by a two-step search feature: 1. the basic (Google-like) small search box, then 2. the pre-determined search filters on the results screen (in a column on the left usually). Even if the latter can help by suggesting additional search avenues, they are not that straightforward to use. They often require knowing beforehand how a field or topic is categorized, giving in some ways less freedom to search (they are pre-determined). They induce the “tree” perspective, rather than the “forest” one (specific versus broad), while both are useful in research. In contrast, advanced search screens offer flexibility, freedom, and some form of control over the search process, once they have been mastered.

No.9 : Then comes the search process, which involves accessing a variety of resources and staying focused, but sufficiently open. Here comes again the essential skill of prioritizing which we discussed (in 4), i.e., selecting the information that best matches the initial requirements and discarding irrelevant ones. If you have clearly defined the initial needs, this task is likely to be more straightforward, returning a limited set of results to process. Again, the importance of acting upstream, i.e., when you identify and set your research needs, which makes it smoother downstream, i.e. when you search and prepare your final research results.

No. 10 : To process the results, selecting and prioritizing search results are essential skills, as is the ability to group, categorize, and classify them. Keeping track of where the most important ones originate is also important, even if it is time-consuming and cumbersome, particularly in extensive research projects. To source them properly, remember the “reproducibility” criteria (in all scientific research, results must be reproducible at a later stage). To help you and your audience process them, contextualize, connect, and justify them in a way that makes them meaningful.

No. 11 : In addition to connecting and documenting your results, you must be able to match them to the requirements, to ensure you have done your job properly and fully addressed the issue/s at stake. Not only at the end. All along the research process, it is necessary to regularly identify any gaps between the initial demand and the desired outcomes. Very often, this comes naturally, as you get stuck in your search. It is time to stop and take a helicopter view of what you have done. Finding any gaps allows you to refocus and continue searching in a more relevant way, making as well sure you address all the aspects of the research. 

No. 12 : The other vital activity, possibly the most important one in our “fake news” and over-abundant poor-quality information world, is to be able to evaluate and assess the information you need to deliver. Concentrate on using “4R” information (Recent – Relevant – Reliable and Rich information). The information you encounter should be questioned endlessly and you should be critical at all times. This is the price to pay to have quality information to take valuable decisions.

No. 13 : Last but not least, delivering ready-to-use, contextualized, connected and meaningful search results is what adds great value to your search efforts. Do not dump your results on your audience; rather, enhance them. Connect and justify them. Finally, success greatly depends on answering the needs you identified at the beginning. Back to competency nr. 1.

These thirteen competencies all require agility, rigor, and assessment skills. This is consistent with the findings of a 2017 Stanford study about digital information evaluation skills[1]. This paper compared three groups: PhDs and information professionals; undergraduate students; and then, fact-checkers. Each was asked to assess the reliability of some websites. The fact-checkers were the most efficient group, as they read laterally, quitting a website after a quick scan and opening a new window to judge the credibility of the original site, moving from site to site without getting lost. The other groups read the websites vertically. Maintaining focus and being agile are two critical aspects of finding and utilizing high-quality information.

[1]              Wineburg, Sam & McGrew, Sarah (2017). Lateral Reading: Reading Less and Learning More When Evaluating Digital Information by Sam Wineburg, Sarah McGrew :: SSRN  Stanford History Education Group Working Paper No. 2017-A1.Published on SSRN [Viewed 27 December 2022].

image from Vectors by AlexandrUA @ https://www.vectorstock.com/royalty-free-vectors/vectors-by_AlexandrUA