Very few former students can claim that they have never written a last -minute research document the day before. AI tools already give new students ways to simulate their papers. NOW, Openai’s new “deep research” tool Seems perfectly designed to help students simulate their path through a term term, unless you are asked to quote sources that do not include Wikipedia.
The new Openai feature, built above its GPT-4O model and published on Sunday, looks like a Google introduced at the end of last year with Gemini 2.0. Google’s “in -depth research” is supposed to generate long form reports in 30 minutes or more, depending on the depth of the requested subject. Google and Openai tools are able to do several internet research while reasoning on the next step to generate a report.
https://www.youtube.com/watch?v=jv-lpisnloo
The in -depth search function is available for Chatppt Pro Subscribers ($ 200 per month) in a new tab under the prompt box. In other words, he does the search on Google for you. Instead of deciding on your own source, the chatbot will do everything that works for you.
For example, Openai asked Chatgpt to look at the iOS and Android adoption rates and the percentage of those who want to learn another language, then find “the markets of which the cat of implementation could develop better”. The Chatbot interface then asked for clarification on several points, such as engaging with a linguistic application was an interest. He looks like questions of questions that your university professor in a level 400 level would ask when presenting your planned research document.
Unlike the Google tool, the deep research of Openai works in a side panel. The chatbot will update you on what it works on. In one of the examples, Openai has shared in his video demo, research on AI seeks surveys on language learning by the country and reports: “I look at surveys and recent reports on the penetration of smartphones and the adoption of the operating system to understand digital trends and the interest of languages on a global scale. “”
Sometimes the bot will cite the sources it looks at, namely Wikipedia. In this example of market studies, the bot has repeatedly examined Wikipedia, just like all lazy first cycle students working on a last minute article the day before. Remember that it is only a scratching internet bot, so it will not accept unlimited books or, on the content, any content locked behind a payment wall.
Once the chatbot generates the report, you can find the complete list of sources in a separate panel. Because it is essentially an automatically geogging machine, AI will probably not have access to the most up -to -date and large -scale surveys of large analysis companies. In the latest article on smartphones adoption rates, the bot cited data from certain sources of trust such as Pew Research. However, he also had various sources, including Wikipedia, Statista, and the magazine of the Gulf Business water industry. This does not mean that the information was inaccurate, but anyone who generates a report is the mercy of the suspicious data and the interpretation by AI of this data.
In one of the examples of Openai on “Air Civil Supersonic – Market Analysis at 10 years” (with title H2 in bold and everything), AI quotes Wikipedia, the web page for Boom Supersonic, the XB-1 developer Jet Plane and the first page of some publications in the industry. In another example, Openai said that his model could find several articles on advanced biology subjects. However, another research document decomposing an episode of a counterpart has torn parts of its information from a fans wiki.
The process takes much more time than any typical Chatppt request. Depending on the depth of a request, it could range from less than 10 minutes to 30 minutes or more. Openai says it comes down to what would normally take humans “several hours”. The company led by Sam Altman said that its in-depth research tool has obtained a higher score than Gemini Thinking, GPT-4O and on The last examination of humanityA AI reference test of IA security center. If the name of this reference seems too close to the Openai and DOMO “End of humanity” of Altman About the future of the AI, it is because the group that made the test A close links with the AI industry and technological billionaires.
OPENAI admitted that in -depth research was “very at an intensity of calculation”. However, he worked on creating a version that would allow more and team subscribers to use it, probably using a less powerful AI model. Most students will not pay $ 200 per month for a cheating tool. Finally, those who cannot spend as much can also be able to pay their Wikipedia cite machine.