The high demand for artificial intelligence (AI) means that tech companies have plentiful opportunities for work but high stakes to deliver quality solutions.
Key Details
- The “AI arms race” began shortly after OpenAI released ChatGPT on November 30, sparking five months of rapid technological innovation and demand for chatbot integration.
- Microsoft, Google, and Baidu have all respectively announced or launched their own chatbot solutions, search engines, and customer service features. At the same time, ChatGPT became the fastest app to reach 100 million users in history.
- As we previously reported, the demand for AI is exceeding the supply, leaving many developers unable to meet the desires of employers and clients who want chatbot-based solutions.
Why It’s Important
With the AI arms race in full swing, the greatest challenges are being left to front-line developers to grapple with the sudden and massive demand for new technologies and solutions. Many companies are struggling to handle the volume of requests and clients, turning many down and focusing on specific projects. At times, the demand exceeds AI capabilities—with clients asking for features that are currently unavailable or are not possible.
Vinay Kumar Sankarapu is the founder and CEO of Arya.ai, a Mumbai-based AI team of roughly 30 researchers and developers that focuses on AI governance for sensitive solutions and industries. The company works with clients around the world, including financial institutions and insurance companies in India, Southeast Asia, and the Middle East.
Sankarapu tells Leaders Media that his company regularly turns down clients, either because it cannot handle the volume or does not want to take on requests that go outside of the direct applications of their AI models.
“This is the problem with not having knowledge about what’s going on behind the scenes. AI can work really well on some complex things, and it can be stupid on simple things. People need to realize what it can solve and cannot solve. Education is lacking today. People who do not know how it works want everything to be done by AI,” says Sankarapu.
Backing Up A Bit
AI integration is the process of creating new processes and solutions that will allow the software to solve new problems or accomplish new tasks, such as installing speech synthesizers or components into the base technology that allows it to interact with users and function as a larger and more complex system.
Developers will underwrite, recommend, and adapt existing AI models and products to new solutions, grappling with which solutions—automation, efficiency, etc.—will improve the work of their clients. This process is the basis of productizing AI and finding ways that the new technology can be profitable and useful.
As Sankarapu notes, the past year has not been the first time AI has been in high demand. Technology goes through cycles of high demand and innovation and has since he launched his company in 2013. The rampant success of ChatGPT has meant that the majority of his product requests are chatbot related, while more seasoned clients seek to use the technology in more specific applications.
“Most clients are asking for conversational assistants as an aid to processes, businesses, and questions. They might have a basic user interface, but why can’t they have a conversational assistant built in? The rest of our requests are product-driven—meaning you have to know how it works to make the most of it. I would say in laymen’s terms that people who do not know AI want a conversation interface for everything, while those with expertise want to use it more strategically,” he says.
Key Takeaways
The rapid proliferation of AI also creates problems with safety concerns and trustworthiness. Companies like Arya.ai build AI models to meet the demands of the market. Still, the demand is moving too quickly, making it difficult for regulators to catch up with the industry and tamp down serious issues that could make the technology dangerous. Arya.ai intimately feels this necessity to deliver trustworthy solutions, working primarily in highly sensitive industries like finance and healthcare.
“We are very much pro-responsible-AI. We operate in very sensitive use cases and industries. We’d rather be cautious than face the wrath of AI, and to enable our products are designed to do that,” says Sankarapu.
The necessity to rapidly productize AI is creating many of the core risks of the moment, particularly with the human element being the most vulnerable and chaotic element in the process. Sankarapu thankfully says the rapid education of the public, employers, and regulators is beginning to catch up to the technology. He expects that improvements are forthcoming.
“Even though there is alot of noise, execution is picking up the pace. People have learned more about AI in the past six months than ever before, and now they are deliberating. People are not finalizing use cases and experimenting. We’ll see alot of rollouts in the next year, followed by a year of crunching as companies roll out more than they can handle and cut off some projects. This is the same cycle we usually see in technology.”