AI Research Showdown Claude Gemini ChatGPT
Anthropic's Claude AI bot, while perhaps not as widely publicized as ChatGPT, Google's Gemini, or Copilot, is gaining traction as a go-to AI for many. It's recognized for being smart, thorough, and offering responses that often feel less robotic and more genuine than many competitors.
Recently, Claude introduced a deep research capability, simply named "Research." This feature, already present in bots from Google and OpenAI, functions like an advanced web search, delivering detailed, lengthy reports instead of brief summaries. It's ideal for research projects or any situation requiring an in-depth exploration of a topic.
These AI models depend on information available on the web. Consequently, their reports may lack comprehensiveness if online material about a specific topic is scarce. However, when resources are available, these deep research tools generally excel at collating information, consistently providing citations and web links to allow users to verify their findings.
Access to Claude's research tool currently requires a paid subscription (starting from $20 a month). Users can select the Research option in the prompt window before entering their query. To evaluate its effectiveness, I compared it with similar features in ChatGPT and Gemini, using subjects I'm familiar with for easier error checking: the Samsung Galaxy S25 Ultra and the filmography of David Lynch.
Researching the Samsung Galaxy S25 Ultra
Claude Research hits a good balance between speed and depth. Credit: Lifehacker
For this test, I used the following prompt for all AI bots:
"I need to know as much as possible about the Samsung Galaxy S25 Ultra. What are its key specs and features? What has changed from the Galaxy S24 Ultra? How does it compare to the other Galaxy S25 models? What kind of users is it suitable for? What does it tell us about the state of the mobile industry in 2025?"
Claude initiated its research promptly and delivered a report in seven minutes, offering an option to view its process. The final 1,200-word document was thorough and accurate, with embedded web links supporting its statements. It highlighted key aspects like AI software and the S Pen downgrade, though some parts of the text were somewhat generic. It was the shortest report among the three.
Gemini was the quickest, producing its report slightly faster than Claude and also generating the most content—over 5,500 words. It first presented a research plan for review and approval, followed by detailed progress updates. The resulting document was highly detailed, including specifics like processor timings and the phone's unveiling date and location. It was accurate and well-referenced, though it frequently veered into Samsung's marketing language.
ChatGPT began by asking clarifying questions and took more than twice as long as Claude and Gemini to produce a report of just over 5,100 words. This bot provides the most detailed insight into its working process, though this is hidden by default. The output was another well-researched, well-formatted, and accurate report with a good selection of details, although Gemini offered slightly more context. Similar to Gemini, ChatGPT tended to lift promotional phrases directly from Samsung's press releases.
Researching David Lynch
Your finished report can be read inside Claude or exported. Credit: Lifehacker
Moving to the works of David Lynch, the prompt was:
"I need a detailed research report into the film and TV works of David Lynch (I don't need to know about his music or art in this case). Which works are considered the most influential? Which are his most celebrated and accessible works? What defines the David Lynch style? What impact has he had on the industry?"
Claude was again the fastest and most concise, delivering just under 1,200 words in a little over five minutes. It accurately identified Lynch's most acclaimed works, explained their significance, and did a good job identifying films and shows influenced by him. While not the most in-depth, Claude provided a solid overview.
Gemini took a couple of minutes longer than Claude to generate nearly 7,000 words on David Lynch, again starting with a research plan. As with the Samsung report, Gemini offered significantly more detail than Claude, including an executive summary, a career timeline (with a table), and a comprehensive, accurate analysis of what "Lynchian" means, complete with quotes from Lynch himself. It served as a very comprehensive starting point for researching Lynch.
ChatGPT took a few more minutes than Gemini, producing a report close to 8,000 words. It followed its pattern of asking clarifying questions initially. The report included a detailed chronology and effectively highlighted Lynch's celebrated and accessible works, as well as some deeper cuts. It accurately summarized what makes Lynch distinctive, though it appeared to rely heavily on a limited number of sources.
The Future of Research
ChatGPT will show you all of its working as it goes. Credit: Lifehacker
Choosing a definitive winner is challenging due to the many variables involved, and preferences for report length can vary. All three chatbots performed well in their research tasks. However, if forced to choose, I would lean towards the Gemini reports for their slightly superior detail and thoughtful organization.
There's undeniable value in AI-powered search, which can produce detailed and seemingly accurate reports in minutes—a task that might take a human researcher hours. When these bots operate in research mode and are compelled to check their work, the likelihood of them fabricating facts should theoretically decrease. In my tests, they appeared to use respectable and authoritative websites as sources.
However, the rise of this technology prompts two critical questions: Firstly, what impact will it have on our own abilities to research, collate information, summarize, and write effectively? Secondly, what is the long-term consequence of users spending more time interacting with AI bots instead of engaging with the open web?
If AI inadvertently leads to the decline of the websites it sources from, these tools will become far less useful. For instance, the bots might be left trying to summarize David Lynch's work based solely on Reddit and Wikipedia, rather than on the insights of professional critics and interviews with industry figures. While this deep research capability is one of the best applications of AI like Claude, Gemini, and ChatGPT, its effectiveness is contingent on the vast library of online content that it simultaneously endangers.
Disclosure: Lifehacker’s parent company, Ziff Davis, filed a lawsuit against OpenAI in April, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.
