‘AI hallucinations are hard to remove completely': Naveen Rao, VP of AI, Databricks

Data and analytics firm Databricks was primed to ride the AI wave as data platform is crucial to train models. The company pioneered 'lakehouse architecture', an open data management system that combines flexibility, cost efficiency and scalability. Post a $10 billion funding round in December that valued the company at $62 billion, the AI company is being keenly watched by investors. But with the dramatic rise in valuations of AI firms comes the pressure to sell AI tools.
In an exclusive interaction with The Hindu, Naveen Rao, VP of AI at Databricks, shared about the enterprise AI market, the AGI moment, and the hype around AI agents.
THB: AI adoption is in full swing even as hallucinations persist. Why is this the case?
Naveen Rao: There are a few reasons. We've gotten more mature on what our application areas are for AI. Basically, we have figured out the use cases where we can tolerate some error. Initially, people tried to apply them to all kinds of areas where precision was required which was a mistake. Secondly, we are adding value in these areas so the models have actually gotten better. When you put information into context either through retrieval or just entering all the required context into the prompt, we see less hallucinations. The models adhere to the information in the prompts with higher fidelity. Over time, we'll continue to work at driving down hallucinations. There can be other checks and balances like having another model judge whether the output is good. That gives us higher precision — it's not 100% but in many cases it's approaching north of 90 — 95% accuracy. But hallucinations are intrinsic to these AI models because LLMs are probabilistic, right? It's an auto regressive, next token prediction. So, it's very hard to remove completely.
THB: Databricks recently announced a $ 100 million partnership with Anthropic. How will this play out in the agentic AI market?
Naveen: We can't have AI agents in companies because there are errors, especially when they're doing multi-step tasks. I do think AI agents are definitely hyped. The definition of an AI agent has shifted to fit the narrative. The original intent of the word was to describe an entity that has agency meaning it can act completely on its own. Now, it has evolved into multi-part systems where multiple LLMs work together to solve a task that usually makes humans faster or more efficient.
Even if AI agents were perfectly behaved, we don't want something that can act completely on its own with no governance. We have built this into Databricks' governance layer for data extended to GenAI. So, whenever an AI agent is built, it has certain access rights and entitlements and it can't go willy-nilly everywhere. In the coming years, as we start to solve these problems, things will become reliable and accurate.
THB: Can we attain AGI with the current level of LLM advances?
Naveen: I mean, whether it's a route to AGI or not will be seen in the technology itself. I personally don't think it is. I don't think that autoregressive loss is the right way to make something that can truly understand causation in a system. Humans learn in a different way. We learn by coming up with a mechanistic understanding of how we can solve a task. So, one thing leads to another thing then another. When I want to solve the same task again, I have a sense of the inputs that cause the causation to the next state. LLMs don't do this. Maybe we'll get there, but the current paradigm of very large pre-training on a huge corpus of unstructured data and then doing some sort of reinforcement learning to modify its behaviour is not what will lead to something that can truly act on its own.
THB: So, you don't believe that AGI is just around the corner?
Naveen: No, I think it's a much harder problem than a lot of people want to give it credit for and we are not at a point where we're close. Yes, we have made huge progress towards autonomous systems that can understand natural language. LLMs have solved natural language, which is a big thing. I don't want to minimize what's been done or their economic impact — LLMs are very useful tools.
THB: How do you compare enterprise AI against the consumer AI market? Do you find it easier to navigate?
Naveen: Enterprises tend to be slower to adopt. They're generally very rational actors, whereas consumers are somewhat irrational. That makes it harder to go after consumer, because it's hard to understand exactly why they're going to buy. Over time the enterprise market will be bigger than the consumer market, I believe. There are really only a few different product surfaces. Search tools is a big one in the consumer segment, like Perplexity or ChatGPT. Image generation and others are really mostly for fun. But I don't know how much people pay for fun. Usually, the novelty wears off unless they're used for business purposes. Whereas in enterprises we see companies really trying to look for an ROI so they're willing to invest a lot because it means something about the company versus their competitors.
THB: What is AI's killer app now?
Naveen: Right now, it's coding. AI tools tend to be effective when they're structured or their output is easily measured. With writing code, you can tell if the code complies pretty easily. Although coding agents hallucinate a lot but most of it is still useful.
THB: What do you think about the view that students should stop studying software engineering?
Naveen: I don't agree with it — someone has to understand how these systems work. Even if code generation is automated, it doesn't mean that the physics of writing software goes away. Somebody will still have to work with the code and check it. We have to understand the basics. I think it's a very poor advice to say we should stop studying computer science altogether.

Try Our AI Features

Explore what Daily8 AI can do for you:

Learn with AIFact CheckingText to SpeechAI Summary

Related Articles

India Today

2 hours ago

• India Today

Building some code competency Department Of Computer Science, Christ, Bengaluru

With its fast-track curriculum and frontier tech focus, the Department of Computer Science at Christ is where innovation meets education and transforms into real-world outcomes No. 1: DEPARTMENT OF COMPUTER SCIENCE, CHRIST (DEEMED TO BE UNIVERSITY), BENGALURU 5 REASONS WHY IT IS THE BEST Christ has partnered with technology firms to design courses in cutting-edge fields like Robotic Process Automation. This ensures both faculty training and real-time industry monitoring of student progress. The Department of Computer Science encourages early exposure to research through its Undergraduate Research Immersion Programme in gamification, immersive technologies etc. It is backed by an annual industry-academia conclave. Through active international partnerships, BCA students can take part in semester-long research exchanges in countries like Germany and France. This builds their cross-cultural competence and exposes them to global best practices. A mandatory fast-track structure enables students to complete advanced coursework early, aligning with NEP 2020's four-year Honours model and preparing them for direct PhD eligibility. With a focus on incubation, the department supports student-driven innovation while training them in high-demand domains like cybersecurity, Gen AI and computer vision—gearing them for both jobs and entrepreneurship.

Scroll.in

5 hours ago

• Scroll.in

DGCA tells Air India to remove three employees for lapses in crew scheduling

The Directorate General of Civil Aviation on Saturday ordered Air India to remove three officials, including a senior executive part of the flight operations department, from all roles related to crew rostering, The Hindu reported, citing an order. The action was initiated for 'systemic failures' and 'multiple violations concerning flight crew being scheduled and operated despite lapses in licensing, rest, and recency requirements', the newspaper quoted the order as stating. The airline was also directed to initiate internal disciplinary proceedings against the three officials within 10 days. An Air India spokesperson confirmed that the directives had been implemented, PTI reported. The spokesperson stated that the company's chief operations officer will also directly oversee the Integrated Operations Control Centre, which is the airline's operational hub responsible for coordinating flight operations, crew rosters and ensuring compliance with safety and regulatory standards. 'Air India is committed to ensuring total adherence to safety protocols and standard practices,' said the spokesperson. According to The Hindu, the alleged violations came to light during a review following the airline's switch to a new crew scheduling software, which was implemented in May last year. #DGCA penalises #AirIndia scheduling and rostering key officials for "repeated and serious violations" and asks for them, including a Divisional VP, to be removed from their posts. Pilots and Crew have long complained about their flying rosters, lack of sufficient rest time &… — barkha dutt (@BDUTT) June 21, 2025 Among those facing disciplinary action is Choorah Singh, the divisional vice president of the Integrated Operations Control Centre, which is the airline's operational hub responsible for coordinating flight operations, crew rosters and ensuring compliance with safety and regulatory standards. The other two persons are Pinky Mittal, the chief manager in the Directorate of Operations, and Payal Arora, who is also involved in rostering, Hindustan Times reported. The directive comes days after Air India's Boeing 787-8 Dreamliner aircraft, which was en route to London's Gatwick airport from Ahmedabad, crashed just 33 seconds after taking off on June 12. This is being viewed as the world's worst aviation disaster in a decade. There were 242 people aboard the aircraft. One passenger survived with ' impact injuries '. Thirty-four persons were also killed on the ground after the plane crashed into the hostel building of the BJ Medical College and Hospital in Ahmedabad, according to Air India. On Wednesday, the airline said that it will reduce its international services that use widebody aircraft by 15% for the 'next few weeks' to ensure operational stability.

The Hindu

6 hours ago

• The Hindu

Bike taxi riders stage silent protest at Vidhan Soudha in Bengaluru

A group of bike taxi riders assembled in front of Vidhan Soudha in Bengaluru for a silent protest on June 21 against the government of Karnataka for not supporting their business. They want the government of Karnataka to frame rules for bike taxis. Bike taxi riders protest in Bengaluru On June 16, app-based bike taxi operations were told to stop operations across Karnataka, following the refusal of the High Court to stay an earlier order mandating the suspension of such services unless the State introduces regulatory norms under the Motor Vehicles Act, 1988. On June 13, a Division Bench of the High Court upheld a single judge order dated April 2, which had directed all bike taxi services to wind up operations within six weeks. The move is believed to have affect thousands of riders and commuters. An executive of a leading bike taxi aggregator platform, speaking on condition of anonymity, claimed that over six lakh individuals in Karnataka depend on bike taxi gigs for their livelihood. 'Most of them earn around ₹35,000 a month, and this ban could push many families into financial distress,' the official claimed. Adi Narayana, president of the Bike Taxi Welfare Association, said many riders are approaching the association's offices seeking clarity on their future. 'Most of our captains have EMIs to pay, families to feed, and rent to cover. This is not just about losing a job, it's about losing the ability to survive. Will the government offer any support?' On June 21, some of the affected bike taxi riders turned up in front of Vidhan Soudha for a silent protest. Some of them had come from as far as Nelamangala, Hoskote, Mandya, and Ramanagara. One of them told The Hindu, 'I work in Peenya industrial estate in Bengaluru. My wife works as a maid. We have two children. I was working part-time for a bike taxi operator to supplement my income. Bengaluru is an expensive city, and I could not afford to take care of my family with my salary.' The bike riders had come together for the protest on their own, and were not led by any organisation. On seeing bike taxi riders gathering in front of Vidhan Soudha, police advised them to disperse saying they could not stage a protest at the site without permission. Some of the protesters were taken into custody while the others were told to disperse, or risk arrest. One bike taxi rider told The Hindu, 'We were not aware that we needed permission to protest in front of Vidhan Soudha in Bengaluru. We will come again. Police advised us to taken permission and protest at Freedom Park in Bengaluru.' Following the ban on their operations, a Rapido spokesperson had said, 'This is a difficult time for our captains, who depend on us for their main source of income. They have played a key role in enabling affordable, last-mile transportation for millions across Karnataka.' The company stated that it is actively engaging with the Karnataka Transport Department to develop a regulatory framework that is not only lawful, but also viable for all stakeholders involved. It stressed that the safety of commuters and the well-being of gig workers will continue to be at the heart of its approach. 'We remain optimistic that ongoing discussions with the authorities will lead to a balanced policy framework, similar to those in other States and Union Territories,' the spokesperson had added.