An artificial intelligence startup has developed Krutrim AI, a generative AI assistant that can converse in over ten languages (Hindi, English, Tamil, Telugu, Malayalam, Bengali, Marathi, Kannada, Gujarati, etc.). Bhavish Aggarwal, also the originator of Ola taxis, established the company. Krutrim AI was developed in-house to guarantee that responses generated by a creative AI tool intended for more than 1.4 billion Indians are one hundred percent contextually pertinent. The development of public-facing artificial intelligence in India has reached a pivotal juncture. Four advantages comprise Krutrim AI.
- Multiple Indian languages are supported.
- The purpose of native experience is to accommodate cultural diversity.
- Trained in tokens and sources of the local language.
- It enables free user interaction on its free version.
Krutrim is currently available in public beta, and a sophisticated version called Krutrim Pro is planned. It was developed by Krutrim SI Designs, a unicorn venture in artificial intelligence led by Bhavish Aggarwal, the inventor of OLAcabs.
Let us now examine how to obtain access to Krutrim AI and compare its performance to the competition’s.
How To Access Krutrim AI?
To access Krutrim AI, you need an OLA taxi-registered phone number or Google account. The detailed procedures are delineated below.
- To log in, navigate to the OLA Krutrim homepage. Enter your mobile phone number or use the Google sign-in option to log in.
- I selected the mobile number option, and an OTP was sent to complete the login process.
- The language of the user interface must be chosen on the following screen.
- Once the user interface language is selected, the conversation space becomes accessible and grants access to Krutrim AI.
How Krutrim AI Works?
Similar to other generative AI chatbots, Krutrim AI works by responding to questions in the conversation field. Large AI models that comprehend prompts in the language used for everyday communication and respond with language similar to humans are known as generative chatbots, like Krutrim AI.
Like a user’s experience with a human assistant, Krutrim AI enables users to respond, modify the details, or add additional components to the response. Despite its technological advancements, Krutrim AI responses still need to be verified by humans because the information provided could be wrong or deceitful. Before acting on such AI-generated output, users are advised to be cautious.
Benefits of Using Krutrim AI
Krutrim AI provides four advantages.
- It is multilingual because it can respond in more than ten Indian languages, which is helpful for users who do not speak English well.
- Cultural context imparts an authentic experience, as Krutrim’s training data was purposefully designed to be culturally sensitive and comprehend the intricacies of the Indian context.
- From the ground up, the development of (AI) startup Krutrim was conducted in House. It is trained on more than two quadrillion language tokens, the greatest number of which represent Indian languages in the landscape of artificial intelligence.
- With a premium version (Krutrim Pro) and more robust knowledge-based capabilities, Krutrim’s primary model provides a free AI chatbot experience.
Krutrim AI vs. ChatGPT vs. Gemini
Well, Krutrim AI will compete with OpenAI’s ChatGPT and Google’s Gemini. Krutrim’s competitors have spent considerable time on the market, so I expect the new entrant to face a formidable challenge. With the following results, eight prompts were used to compare the task execution and knowledge-based capabilities of Krutrim AI, ChatGPT, and Gemini.
Grammatical Correction
All chatbots rectified the sentence; however, Gemini received additional credit for the clarifications.
Krutrim scores one, ChatGPT scores one, Gemini scores two.
Rhetorical Questions
Regarding addressing global poverty, Krutrim AI, ChatGPT, and Gemini all respond thoroughly. However, it would have been more accurate to present certain significant aspects, such as social security programs, infrastructure development, and others, in their distinct sections rather than combine them. On the other hand, Krutrim could have incorporated principles number five and eight. As a result, ChatGPT from OpenAI emerges victorious in this round.
ChatGPT scores one.
Creative Writing
To begin with, all three scenarios were excellent. Krutrim’s creation, however, had a more uplifting feel than one that was funny. ChatGPT performed more effectively. However, Gemini captured it the most. The most humorous aspect was the concept of an ant becoming agitated and flailing its fist at an elephant, followed by the remainder of the tale.
Gemini scores one.
Comprehensive Comparison
Krutrim performed exceptionally poorly, adhering to the allotted 200-word limit with only 51 words. Gemini and ChatGPT were nearby, but the latter provided more lucid data, which is helpful when deciding.
Gemini scores one.
Telling a Joke
Of the three, Krutrim was the least funny. Although ChatGPT was commendable, the second portion of the jest lacked coherence. When the cat responded cynically to the dog’s attempt to engage with it, Gemini once more delivered his finest performance.
Gemini scores one.
Opinion Questions
Krutrim’s response was once more unnecessarily concise, devoid of the requisite details that a reader might demand. As a result of its data exclusion, ChatGPT was limited to responding exclusively to its training data. Through its internet connectivity, Gemini provided hyperlinks to several supplementary reading materials. Nevertheless, the response generated by Gemini needed to catch up to the quality of ChatGPT’s output.
ChatGPT scores one.
Drafting Skills
Google is undoubtedly the authority on advertisements, which is abundantly obvious in Gemini’s outstanding performance. Although the responses from Krutrim and ChatGPT weren’t awful, Gemini’s options for the advertisement, headline, description, and call-to-action were in a different league.
Gemini scores one.
Subjective Questions
Gemini fell empty in this manner for the very first time despite its continual internet connectivity. ChatGPT performed best. In addition to the standard response provided by ChatGPT, Krutrim enumerated several well-known individuals who drew inspiration from the Indian peace icon.
Krutrim scores one.
Our fleeting confrontation between these three generative AI assistants has ended. The final scorecard is as follows.
| Prompt Type | Krutrim | ChatGPT | Gemini |
|---|---|---|---|
| Grammatical Correction | 1 | 1 | 2 |
| Rhetorical Question | 0 | 1 | 0 |
| Comprehensive Comparison | 0 | 0 | 1 |
| Creative Writing | 0 | 0 | 1 |
| Telling a Joke | 0 | 0 | 1 |
| Opinion Questions | 0 | 1 | 0 |
| Drafting Skills | 0 | 0 | 1 |
| Subjective Questions | 1 | 0 | 0 |
| Final Score | 2 | 3 | 6 |
Gemini from Google emerged as the unequivocal victor, distinguishing itself through its comprehensive and logically organized responses. ChatGPT performed admirably and was able to maintain its ground. Conversely, Krutrim AI came last, indicating that its knowledge-based functionalities have considerable room for enhancement.
At the time of their beta releases, ChatGPT and Gemini (formerly Bard) were, nonetheless, the subject of jests on digital platforms. This historical account underscores the potential for Krutrim AI to become a leading regional generative AI assistant, given sufficient user feedback and a steadfast dedication to its development by OLA’s artificial intelligence firm. Considering cultural and regional factors, its beta release has demonstrated its value to Indian citizens and is positioned to impact enterprises substantially. Let us examine the opinions of the initial evaluators regarding Krutrim.
Krutrim AI Public Response
The public reaction to Ola Krutrim AI is, at best, inconclusive, as evidenced by user tweets on the social media platform Twitter, which vary in quality according to the effectiveness of Krutrim’s responses. In response to questions about local construction, cultural apparel trends, and the translation of scriptures and Indian language materials into English, among other things, Ola Krutrim AI has received favorable public feedback. However, it did demonstrate restricted functionality in providing the accurate date when prompted. However, it is important to remember that this is still a beta version and a more sophisticated model is currently in the works.
Understanding Indian languages, region-specific information, and cultural context is a strength of Krutrim AI. Although much progress remains to be made, there are instances where things are becoming more favorable. As is the case with any AI tool, be cautious and verify the information provided.
Author Opinion
Krutrim executes the incorporation of public feedback satisfactorily. My experience with it was similar to that of other generative AI assistants, such as ChatGPT and Bard (now Gemini), despite it still having some raw aspects. At present, Gemini is the most impressive for an international audience. You can use Microsoft Copilot to access advanced OpenAI GPT models for free during off-peak hours.
Advice When Using AI
Consider the following three points when using AI chatbots such as Krutrim, ChatGPT, and Gemini.
- Answers must be validated, as AI may experience hallucinations.
- Real-time data access and language comprehension capabilities are subject to constraints. If the chatbot is incapable of comprehending the subtleties of language, the response may not be the most accurate or completely wrong.
- To safeguard one’s privacy, refrain from disclosing any personally identifiable information.
Also, Check:
