Genei: Can A Summariser Help You Overcome Information Overload?
Summarisers have been around for some time, and their use cases have matured significantly with technological advancements. Among the numerous summariser apps available, Genei stands out as a cutting-edge tool that provides a more personalised and efficient summarisation experience.
With an app such as Genei, users can tackle more complex pieces of writing and tailor summaries to their preferences. In this review, we will discuss the good, the bad, and the ugly aspects of Genei while highlighting the impact of the latest technology on its functionality.
Genei is an AI-powered summarisation and research tool that can summarise a webpage or uploaded PDF document. The obvious benefit of using Genei is saving time by reading a summary rather than an entire document. This can be particularly handy if you are a student or academic and spend hours trawling through research papers only to find that it doesn't have the information you are after.
This summarisation service does come with a cost. The basic plan will set you back £9.99 per month, while the Pro plan costs £29.99 per month. However, you can save a few pounds by opting for a yearly plan, and academics can receive a 40% discount.
From the pricing alone, you can see Genei is targeting that academic market, whether they're students or post-grad researchers. Note that you need to enter an academic email address for verification to receive the academic discount.
There is a 14-day free trial when you sign up; however, you must still enter your credit card information to receive the free trial.
The interface is clean and simple, allowing you to upload a PDF or paste a link to the web page you want summarised. There is also the ability to read scanned documents, which is a plus as many older research papers are often scanned from the original print.
The drag and drop PDF screen
We asked Genei to summarise two web pages. The first is the Nielsen Norman Group article about ChatGPT increasing business professionals' productivity and work quality.
On the initial upload, Genei will display a text version of the document so that you can read through the entire article (should you wish) and make any highlights or copy text. Genei also scans the paper looking for headings or sections. You'll see this in the Summary section. You can then expand each section to read specific summaries, or you can elect to "Compute all" to see the entire summarisation.
Output Summary from Nielsen Norman Group article
ChatGPT Lifts Business Professionals' Productivity and Improves Work Quality
- The recent introduction of ChatGPT and similar generative AI tools has led to concerns about their accuracy and potential to produce misleading or false information.
- While it is possible to trick these tools into generating incorrect or harmful output, this is not necessarily an inherent flaw of the technology.
- The key question is whether the tool produces good results when used as intended.
- A new research study explores the potential benefits of using ChatGPT with human oversight, suggesting that with proper editing and corrections, the tool can generate high-quality output.
- The section discusses an empirical study conducted by Shakked Noy and Whitney Zhang from MIT, which involved 444 business professionals from various fields using ChatGPT to write business documents.
- Participants were randomly assigned to use ChatGPT or write without AI assistance.
- The study found that ChatGPT has great usability for first-time users and that the quality of work produced with or without AI assistance was graded by independent evaluators.
- The study highlights the importance of assessing the quality of work produced with a tool in UX studies.
Results: Faster Work, Better Results
- This section discusses a study that examined the use of ChatGPT by business professionals.
- The study found that using ChatGPT resulted in increased productivity and higher quality deliverables.
- The productivity improvement was 59%, with an effect size of 0.83 standard deviations, while the quality improvement was 0.7 on a 7-point scale, with an effect size of 0.45 standard deviations.
- These improvements were statistically significant and were achieved even though most participants had no prior experience with ChatGPT.
- The section concludes that long-term improvements are likely to be much bigger as users discover better ways of using the tool and adapt their work styles accordingly.
Why Better Performance with ChatGPT
- The study found that using ChatGPT reduced skill inequalities and changed the way users allocated their time in the writing process.
- Participants spent less time generating rough drafts and more time polishing the final result, resulting in a reduction in task time and potentially contributing to higher-rated document quality.
- The main contribution of ChatGPT was saving users substantial time on the production of rough text.
- The study did not fully answer why business professionals performed better with ChatGPT, but some interesting insights emerged.
- The authors of the study on ChatGPT are praised for providing empirical data on real business professionals using the technology for realistic tasks.
- However, the study has some weaknesses, including a limited range of professions studied and a lack of qualitative insight into user behavior.
- To fully understand the impact of AI on business professionals, data from a wider range of jobs and larger-scoped tasks is needed.
- Additionally, the estimates of how users divided their time were based on self-reported data, which is weak in UX studies.
- Future research should address these limitations.
- The use of generative AI like ChatGPT can result in biased or false output, but when combined with skilled humans, the results can be superior.
- AI can augment human intellect rather than replace it, as envisioned by Doug Engelbart for advanced user interfaces.
- Therefore, when considering the introduction of AI tools in business, it is important to explore ways in which AI and human professionals can work together.
- The provided section only contains a reference to a working paper by Shakked Noy and Whitney Zhang titled "Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence" which can be accessed through a link to the MIT Economics Department website.
- No further information or insights are provided in this section.
The 2,000-word article has now been distilled into a 600-word summary. The length of the summaries can be customised in the settings, with short, standard, and long available. This example was created on the standard setting.
You may have noticed two bullet points highlighted in red in the above summary. After reading the entire article and the summary, we concluded that these two points are incorrect. Two out of 27 bullet points isn't a bad result and occurred because the software didn't quite understand the article.
We asked Genei to summarise the next web page from Open AI's blog.
Output summary from Open AI's blog
- The section describes the ChatGPT model, which is capable of conversational dialogue and can answer follow-up questions, admit mistakes, and reject inappropriate requests.
- It is a sibling model to InstructGPT, which is trained to provide detailed responses to prompts.
- ChatGPT is available for free usage during the research preview.
- The section also provides various examples of ChatGPT's capabilities, including debugging code, writing formal letters, and discussing Fermat's Little Theorem.
- Fermat's Little Theorem is a useful tool in cryptography for performing modular exponentiation efficiently and generating private keys from public keys.
- Finally, the section includes a sample introduction letter that ChatGPT can help make more formal.
- The section includes various topics, including a letter template for introducing oneself to a new neighbor, troubleshooting a code issue, and discussing the development of ChatGPT, an AI system trained using reinforcement learning from human feedback (RLHF).
- The section also highlights the importance of providing user feedback to improve the model's performance and mitigate potential risks.
- The ChatGPT model was fine-tuned using a combination of supervised fine-tuning and RLHF, and its safety mitigations were informed by earlier model deployments.
- The OpenAI team encourages users to provide feedback on problematic outputs and false positives/negatives through the ChatGPT interface and the feedback form.
- The section also includes a disclaimer that the AI assistant cannot provide violent or gory content and that bullying is never acceptable.
From the output, we can see that Genei has trouble summarising content that doesn't have clear headings or sections. The summaries often start with "The section…" which is vague and unhelpful. However, once you click on the summary, it will highlight the exact section it refers to in the text.
Genei can also help you generate notes and create content using their prompts library or your own. This includes paraphrasing text, creating essay outlines, writing social media posts, developing blog posts, and more.
Genei Options Menu
Genei enables users to research much faster by summarising and extracting keywords from academic papers or articles. Users can easily upload content, and the app will analyse it and provide abstract summaries and section-by-section breakdowns. This level of detail and organisation makes the app indispensable for researchers and professionals who need to digest the main points of a paper or article quickly.
We also asked ChatGPT to summarise the same articles for comparison purposes.
ChatGPT Output Summary Comparison
Example 1 summary from ChatGPT:
- OpenAI has developed a large language model called ChatGPT based on the GPT-3 architecture.
- ChatGPT is capable of conversing with humans on a wide range of topics and has been trained on a massive amount of text from the internet.
- The training data used to create ChatGPT included a diverse range of sources, including books, websites, and social media platforms.
- ChatGPT has several potential applications, including in customer service, education, and mental health.
- The development of ChatGPT raises ethical concerns about its potential impact on employment, privacy, and bias.
- OpenAI has implemented several measures to address these concerns, such as limiting access to the model and implementing safeguards to prevent its misuse.
- ChatGPT can also be used for creative applications, such as generating poetry and music.
- The development of ChatGPT is part of a broader trend in the field of artificial intelligence toward creating more advanced language models.
- The article discusses the challenges and limitations of language models, including their tendency to perpetuate biases and their reliance on large amounts of data.
- Overall, ChatGPT represents a significant advancement in the field of natural language processing and has the potential to transform many aspects of human life.
Example 2 summary from ChatGPT:
The article discusses how the large language model ChatGPT, developed by OpenAI, can be used to increase productivity in various tasks such as research, scheduling, note-taking, generating ideas, and providing feedback on writing. The article provides tips on how to use ChatGPT effectively and how it can be integrated with existing productivity tools. However, the development of language models raises ethical concerns about their impact on employment and privacy. While ChatGPT can augment human capabilities, it is not a replacement for human expertise. As language models continue to improve, they have the potential to transform many aspects of work and life.
How does Genei compare to ChatGPT? ITS BETTER.
To be honest, ChatGPT did a terrible job summarising these articles. While the summaries looked factually correct, they were very general, and the content did not seem to originate from the article but from public knowledge of the topic.
For example, the first summary highlights that ChatGPT can be used for creative applications, such as generating poetry and music. Yes, this statement is factually correct; however, this was never discussed in the article.
The issue with ChatGPT is it cannot evaluate between when to only use the source material provided or statisically generated response. It wants to guess at your intentions, rather than provide contextually factual statements. So you must prompt it a few times carefully to ensure it analyses its own response so it can check for factuality. Remember it's a tool suited to a specific use case, chat, and not a thinking machine.
ChatGPT also doesn't have an option out of the box to upload PDFs for summarisation, so Genei edges out ChatGPT in that feature. Overall, we recommend skipping on custom summariser prompts using ChatGPT and instead use a summariser like Genei in your workflow.
The Bad - NO TL;DR
Genie is great at summarising an article section by section. Where it falls short is providing an overall summary of the article. It would be helpful to receive a short three-sentence summary of an article first to decide if it is worthwhile summarising the entire thing in sections.
Further, if your article lacks structure in headings and sections, Genei's summaries become sloppy and harder to follow.
Genei's summaries are imperfect, and relying on them too heavily can detract from the whole experience of reading and understanding a piece of writing. Although academic papers may benefit from summarisation, using these tools for all articles may result in a less enjoyable and nuanced reading experience, and a mistaken self belief you have sufficiently studied a topic. Reading a quality article in it's entirety you will fully appreciate the author's intent, the context, and the intricacies of their work.
Who is Genei for?
If you deal heavily in research, whether for school, academic research, or professional writing, Genei is a great tool to use to help overcome information overload. It helps save time and allows you to understand complex pieces of writing quickly. Also it can expand your sources by valuating a set of articles to see which one is most relevent for your further study.
If you want to engage with the content thoroughly and appreciate the intricacies of a piece of writing, taking the time to read the entire article is still necessary. Ultimately, it is essential to strike the right balance between efficiency and depth of understanding when using summariser apps like Genei.