A new report from the US digital marketing company Legal Guardian Digital has attracted attention when ranking Perplexity AI as the most reliable AI chatbot for daily work tasks.
This result shows that Perplexity surpasses many famous competitors such as OpenAI with ChatGPT, Google with Google Gemini or Anthropic with Claude.
According to research, AI chatbots are evaluated based on many criteria including the rate of creating misinformation, customer satisfaction, consistency in feedback and system stability.
The report was released in the context of AI becoming increasingly widely used in the workplace, when about a quarter of American workers are said to regularly use AI tools to support work.
The results show that Perplexity AI has the lowest rate of "hallucinatory" and untrue information in the large chatbot group, only at 13%, significantly lower than the industry average of 22%.
This platform also achieves a 100% continuous operation rate, meaning no service interruption time is recorded throughout the testing process.
Second place is Grok, developed by billionaire Elon Musk's AI company. Grok recorded a rate of creating incorrect information at 15% and also achieved perfect operating time.
Meanwhile, DeepSeek ranked third with an "illusion" rate of 14% and operating time of 99.52%.
The most surprising thing in the ranking is the position of ChatGPT. Although it is the most popular AI chatbot in the world today, ChatGPT only ranks sixth in overall reliability.
The report said that this chatbot gives inaccurate feedback in about 30% of cases, twice as high as DeepSeek's error rate.
However, ChatGPT still received positive reviews from users with a satisfaction score of 4.7/5.
In the lower positions, Google Gemini ranked eighth, while Meta with Meta AI ranked ninth.
Claude of Anthropic ranked seventh and is said to have had more operating incidents than competitors.
The research team said that they combined many factors to calculate the overall reliability score for each chatbot.
In which, important factors include the frequency of returning incorrect information, stable responsiveness to many types of questions, user reviews and system stability.
Perplexity AI achieved the highest trust score with 85/100 points. Grok followed with 79 points, while DeepSeek achieved 76 points. ChatGPT only received 50 points, while Google Gemini achieved 41 points.
The report also mentions the service price. Perplexity AI is said to be priced at about 40 USD per month, Grok is about 30 USD per month, and DeepSeek is currently provided for free even though it is still in the most reliable chatbot group.
Mr. Austin Hunt - CEO of Legal Guardian Digital, said that many people default to ChatGPT as the most reliable chatbot because of its pioneering advantage and global popularity.
However, in terms of stability and accuracy in work, smaller platforms like Perplexity or Grok are showing better efficiency.
