Google is upgraded GeminiHis chatbot platform, with the possibility of “reasoning” through a research problem and compiling a complete report.
The updated gemini offers a functionality called “deep research” which, according to Google, uses “advanced reasoning” and “long context capacities” to generate research memories. Memoirs are presented in Gemini applications and can be exported to Google Docs for additional modification.
For the moment, deep research is exclusive to Gemini Advanced, a more sophisticated version of Gemini Google One Ai Premium Planat a price of $ 20 per month.
Deep
Google says that in -depth research can analyze the relevant information for a request through the web in the name of a user, acting as a kind of research assistant. The results are organized in summaries in thesis and associated with links to the original source material.
Here’s how it works:
- A user writes a question.
- In -depth research creates a “Search plan in several steps” so that the user revises or approves.
- Once the user has been approved, Deep Research refines his analysis in a few minutes – search, record potentially interesting information, then start a new search according to what it is
learned. - The process is repeated several times, and once finished, deep research generates a report of the main conclusions.
In -depth research is only available in English on Desktop and on the mobile web to start. Users can access it by selecting the “Gemini 1.5 pro With in -depth research ”in the drop -down menu of the model. Google says it will happen to Gemini mobile applications at the beginning of 2025.
Google’s new deep research feature looks like one of the most appropriate “Google-Y” uses of AI to date, and is quite impressive.
I had a little access and he made very good initial reports on almost all the subjects. The pay walls around academic sources put some limits. pic.twitter.com/dwsqr6akgz
– Ethan Mollick (@emollick) December 11, 2024
“We have built a new agentic system that uses Google’s expertise to find relevant information on the web to direct navigation and gem search,” wrote David Citron, product director of Gemini applications, in a blog article provided in Techcrunch. “An in -depth research saves you hours.”
But the functionality raises all kinds of thorny ethical questions.
Potential damage
Put aside the fact that the whole AI make mistakes And hallucinate (Everyone remembers glue pizza?), A technology like in -depth research could have serious consequences for education.
In a recent spur In the New York Times, Jessica Grose wrote about how students are counting more and more about a generative AI to outsource brainstorming and writing. These students, she said, may lose the ability to think critically and overcome their frustration with tasks that cannot be easily completed.
At least one study has linked a strong use of chatgpt among students at higher levels of procrastination, memory loss and a means of lower points.
In -depth research also threatens to harm – in the monetary sense, that is to say – the publishers from which he gets to his information. By scraping information from websites and compiling it in memories, in -depth research could deprive these precious advertising income sites.
The impact of IA previewThe summaries generated by AIM provides Google for certain Google search requests, on publishers can be revealing of what will happen, if in -depth research takes off.
According to At a source, publishers have seen a decrease of 5% to 10% of traffic from research since the IA previews at the start of this year. On the income side, an expert quoted by the New York Post estimated The fact that the previews generated by the AII could cause more than $ 2 billion in losses for publishers.
Google often says that it works in close collaboration with the publisher’s partners, respects payment walls and allows websites to block its scratch of AI at the field. But the sockets are often faced with a dilemma: Allow scratching or losing visibility in Google research (at least for the moment).
Google claims that in -depth research could “connect users to relevant websites that they would not have found otherwise so that they can dive deeper to find out more”. We will have to see if the functionality keeps this promise to discover or if it diverts the views of the wider web.
Gémini 2.0 Flash
In -depth research is not the only new ability to come to Gemini.
From today, free and paid Gemeni users will have access to Gémini 2.0 FlashThe new model of Google lighthouse. To be precise, it is an experimental version of 2.0 flash optimized for the cat – the full version will arrive in January.
Google says that 2.0 flash, which can be selected from the Gemini drop -down model on the desktop and mobile web (but not yet mobile applications), should offer better performance on a certain number of tasks – and faster responses.
However, the company warns certain gemini features “will not be compatible with (the) model in its experimental state”. He did not say exactly what features.