Drums | Where’s Watson? IBM’s Tarun Chopra on his company’s AI strategy: past, present and future

AI For Business

IBM is a pioneer in demonstrating the power of artificial intelligence to the world. But as talk of the burgeoning “AI revolution” dominates the media landscape, the tech giant has conspicuously disappeared from the headlines. So where the heck is IBM, and what happened to Watson?

In 2010, IBM released Watson, a system that uses natural language processing (NLP) and machine learning to answer questions. The following year, Watson, named after IBM founder Thomas J. Watson, won Jeopardy!. Play against two former (human) champions.

Watson’s Crisis! This victory cemented both Systems and IBM in artificial intelligence (AI) history. For the first time, computers have clearly outperformed the human brain in answering difficult questions accurately. This isn’t the first time an AI-powered computer system designed by IBM has surpassed humans in one of our most cherished intellectual domains. In 1997, the company’s Deep Blue supercomputer defeated chess grandmaster Gary Kasparov.

Nearly two years after Watson’s historic Dangerous Incident! With the victory, Memorial Sloan-Kettering Cancer Center in New York City is using computers to help doctors make more accurate diagnoses and recommend more individualized treatment plans. Integrate the system and make Watson the first AI model with a commercial application.

In other words, IBM quickly established itself as an AI trailblazer. The highly publicized Deep Blue and Watson moment was akin to a moon landing in a field that was still burgeoning at the time. But where has the company been since then?

Since the November 2022 release of ChatGPT, an AI model designed to respond to text-based prompts in natural language developed by OpenAI, the media industry has been buzzing about AI. Every week, new AI-powered products are released, and experts issue new dire warnings, reigniting public interest (and fear) in this rapidly advancing technology. . We are currently living in the early stages of what is called the “AI Revolution”. It may be exaggerated, but it may not be.

Google, Microsoft, Meta, and other tech giants are spending billions of dollars to gain a foothold in the burgeoning AI industry. But where is IBM?

To find out, The Drum spoke with Tarun Chopra, vice president of product management, data and AI at IBM.

It’s been 13 years since Watson went on sale and 10 years since the computer system was integrated into its first commercial use (by Memorial Sloan Kettering Cancer Center in New York City). What has IBM’s AI strategy been like since then, and where has the company been primarily focused?

Based on machine learning technology, Watson takes data as input and builds a model. Then, in typical corporate fashion, he worked with thousands of clients as he implemented AI. [business] landscape. And we learned a lot. Producing these capabilities in a large enterprise environment is not so easy for one simple reason. There are many factors involved in mainstream large-scale production, including size, corporate regulation, corporate mission, privatization, return on investment, and cost.

This trip was so fascinating. It went at a pace we didn’t expect. That is why we embarked on what we once called “Trusted AI” or “AI Governance”. Because the customer said to us, “It’s okay if it’s a priority.” [AI], must be explainable to shareholders. It should not be like a black box. As such, there have been many innovations in AI explainability. IBM put a lot of effort into this. Because I’ve found that just creating something and publishing it doesn’t work. Customers need a variety of tools for productive deployment. [AI] integrated into their business model. I mean, we were helping bring AI to market, but it wasn’t sexy, so it wasn’t talked about much. It was more like going all out to help a customer with her AI implementation.

Another thing we’ve learned is that it’s not enough to just provide a bespoke AI solution for everyone. Broadly applicable applications are needed to drive adoption.

The release of ChatGPT by OpenAI changed the way most laymen think about AI in many important ways. How has IBM positioned his AI products and services in-house since then?

What ChatGPT really emphasized was the value of this new technology, called foundational models, which allows you to apply entire models to other models instead of relying solely on the model’s data. This opens the door to a more mainstream audience. Because you don’t have to be an expert to use this technology. Just send the prompt. You can add proms and other things to find answers or tweak. However, most companies are not going to adopt this technology because it is still a huge black box. This presents an opportunity for his IBM. We have his 10 years of experience learning how to productize AI into the enterprise, and are now able to apply it to help clients deploy foundational models.

It takes a lot of effort, skill and governance to help large enterprise clients adopt and scale AI. Mass scale of AI is a hot topic in the world right now. Even with all the hype removed, it still has a long way to go to reach that goal. But I think what IBM has learned over the last decade or so is very useful to us and could help us move the needle towards scale.

In 2011, IBM made history when its Watson computer system defeated two top (human) Jeopardys. champions. What did that moment mean to you, working closely with AI?

One of the things it emphasized for me is that what is new today will be old tomorrow.Fun to be in this field if you like that fast pace [technological] change.

It was also appealing to be able to continuously see customer interest, as it can be hard to imagine what you can actually do. [a new technology]. When the Internet was invented, people had no idea what to do with it until Netscape came along. The same thing is happening now with AI, first with the Watson crisis.moment, then [Google’s] In the Deep Blue Chess moment, and now the ChatGPT moment, the average consumer can think, “I can do this with this, or I can do that with that.” But unless you really understand the killer apps and how to make these features available to consumers, mass adoption is difficult.

So for me they are [historic AI] Moments followed, reflecting the customer’s interest again, but the same questions kept popping up. “Then can you provide it?” [technology] What impact will it have on the company’s day-to-day environment? That’s a whole different job.

Recent advances in AI have sent shockwaves of anxiety across society, primarily around unemployment and survival risks. Are there any common misconceptions about AI and its perceived risks? Brands like IBM communicate about AI in a way that presents it not as something to be feared, but as something that legitimately benefits humanity. What do you think about the challenge of

For me, AI will bring enormous benefits to society. It actually presupposes the business problem you’re trying to solve. Can AI be deployed to make humans more productive? Ultimately, productivity is what counts. I believe in the power of AI to extend human capabilities and knowledge. See my own field. AI can be used to create better software. Just because robots will write software tomorrow doesn’t mean they won’t need anything. [human] programmer. It’s more like, “How can we improve the overall quality and productivity of what we do?” [using AI]. ”

If used correctly and communicating in the right way to consumers, governments and markets, AI could be an extension of what we do today.All technology in history is actually meant to scale [human capabilities]. right? AI is the next big evolution in that expansion. That’s my take and it’s based on what’s happening in the market today.

What do you think are the biggest opportunities and challenges that AI presents to marketers today?

I think the biggest opportunity is integrated storytelling. [marketers] AI can be leveraged to ingest input from hundreds of different sources that were not possible before.

The big challenge for marketers isn’t believing all the hype, but knowing how to separate fact from fiction. They need to be able to do their own homework and know the facts. We also have to be careful about spreading misinformation through AI. If AI models start hallucinating, your company might do the same. As the digital space becomes more inundated with AI, it becomes increasingly difficult to separate fact from fiction. This is where governance becomes very interesting and very important.

What’s next for IBM’s AI research?

ChatGPT brought the functionality of the foundational model to the world. Over the next 18-24 months, [at IBM] We will focus on features and models applicable to different industries. A large language model is just one of his features of the underlying model. We are working with a variety of organizations, including NASA and chemical companies, to expand the scope and capabilities of our foundational models. We also work with clients to help them understand, manage, productize, and cost-effectively implement the capabilities of their underlying models.

given a short period of time. [AI] The universe is moving very fast. No one knows what will happen in 10 years.But from a product leader’s perspective, these [projects] I look forward to working on this. At the same time, as I said earlier, in terms of understanding what customers really need to productize these capabilities in their environment, we’ve taken his 10 years of learning from Watson. take advantage of

To learn more about the latest in AI, Web3, and other cutting-edge technologies, sign up for The Emerging Tech Briefing newsletter here.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *