Artificial intelligence has transitioned from a mere bragging right to a proven differentiator for organizations across the country. IDC predicts a 50.1 percent CAGR for global spending on AI, reaching USD 57.6 billion by the year 2021.
With the enterprise realising that AI is no rocket science and roping in the right talent is not impossible, an increasing number of organizations in India are tiptoeing into the AI space. CIO India deep dives into what drives AI adoption, the roadblocks in the AI journey, and why an industry-academia collaboration is of utmost importance today.
Data – the prime mover behind the AI drive
For AI to really work its magic, what’s needed most is data – humongous amounts of it. A study published in the Scientific American reveals that each year, the amount of data we produce doubles. Essentially, we’re currently perched on a treasure trove of usable data.
Chiranjib Bhattacharyya, Professor - Department of Computer Science and Automation at the Indian Institute of Science points out that the most important drivers for AI and machine learning is data. “Government sharing data could enable data liquidity. And data liquidity has the potential to generate USD 3 trillion,” he says.
It’s true the Indian government allocated USD 480 million to develop emerging technologies like AI, ML and IoT. However, the government not announcing tax breaks for companies engaging in R&D activities is not a deterrent for Indian companies.
“Fostering AI growth doesn't necessarily have to be through incentivization or tax breaks. The government can also drive AI adoption by providing data. If startups and tech companies have access to that data, it can prove to be the lifeblood for driving AI," reveals Shrirang Karandikar, GM - Projects and Technology, Shell India.
Barriers in AI – how industry-academia symbiosis and learning makes the difference
Now that it’s known data is the foundation upon which AI is built, for everyone to draw benefits out of it, it is imperative for the industry to collaborate with academicians to increase AI proliferation.
Academicians have the in-depth know how and the bandwidth to invest in research and development – something not all organizations can afford to indulge in. But what can the industry do for the academia?
Bhattacharyya reveals that in most cases academicians do not have the data, but the industry does. "Students are waking up to this reality and are willing to go to companies, although it's not the sort the job they are looking for," he says.
Another stumbling block in the AI journey is the amount of compute needed. Karandikar of Shell India says that the more well-known model of AI is one in which a lot of data is used to train the model. This is the part that requires a lot of data and compute. And the amount of compute required for training the data could otherwise be used for operations.
"If the amount of compute we require for training is going to be as much as the amount we would use for operations, there really is no benefit from it," he adds.
Rajshekar Behar, marketing leader at Julia Computing – a rapidly rising startup specializing in AI solutions, believes that investing in learning holds the key to AI growth.
"The major barrier I see is that the gap between haves and have-nots keeps increasing. One way to tackle this problem is to invest in the infrastructure of learning. The skillset required for AI is not rocket science, so colleges need to include AI in their curriculum," he adds.
Questions around ethos and people losing jobs to AI: in Bhattacharyya's point of view, AI could pose a threat to repetitive jobs - for instance call centres. He also believes that India is well-poised to create mass employment in AI. The professor also expressed the need for a legal framework pertaining to AI.
http://www.cio.in
With the enterprise realising that AI is no rocket science and roping in the right talent is not impossible, an increasing number of organizations in India are tiptoeing into the AI space. CIO India deep dives into what drives AI adoption, the roadblocks in the AI journey, and why an industry-academia collaboration is of utmost importance today.
Data – the prime mover behind the AI drive
For AI to really work its magic, what’s needed most is data – humongous amounts of it. A study published in the Scientific American reveals that each year, the amount of data we produce doubles. Essentially, we’re currently perched on a treasure trove of usable data.
Chiranjib Bhattacharyya, Professor - Department of Computer Science and Automation at the Indian Institute of Science points out that the most important drivers for AI and machine learning is data. “Government sharing data could enable data liquidity. And data liquidity has the potential to generate USD 3 trillion,” he says.
It’s true the Indian government allocated USD 480 million to develop emerging technologies like AI, ML and IoT. However, the government not announcing tax breaks for companies engaging in R&D activities is not a deterrent for Indian companies.
“Fostering AI growth doesn't necessarily have to be through incentivization or tax breaks. The government can also drive AI adoption by providing data. If startups and tech companies have access to that data, it can prove to be the lifeblood for driving AI," reveals Shrirang Karandikar, GM - Projects and Technology, Shell India.
Barriers in AI – how industry-academia symbiosis and learning makes the difference
Now that it’s known data is the foundation upon which AI is built, for everyone to draw benefits out of it, it is imperative for the industry to collaborate with academicians to increase AI proliferation.
Academicians have the in-depth know how and the bandwidth to invest in research and development – something not all organizations can afford to indulge in. But what can the industry do for the academia?
Bhattacharyya reveals that in most cases academicians do not have the data, but the industry does. "Students are waking up to this reality and are willing to go to companies, although it's not the sort the job they are looking for," he says.
Another stumbling block in the AI journey is the amount of compute needed. Karandikar of Shell India says that the more well-known model of AI is one in which a lot of data is used to train the model. This is the part that requires a lot of data and compute. And the amount of compute required for training the data could otherwise be used for operations.
"If the amount of compute we require for training is going to be as much as the amount we would use for operations, there really is no benefit from it," he adds.
Rajshekar Behar, marketing leader at Julia Computing – a rapidly rising startup specializing in AI solutions, believes that investing in learning holds the key to AI growth.
"The major barrier I see is that the gap between haves and have-nots keeps increasing. One way to tackle this problem is to invest in the infrastructure of learning. The skillset required for AI is not rocket science, so colleges need to include AI in their curriculum," he adds.
Questions around ethos and people losing jobs to AI: in Bhattacharyya's point of view, AI could pose a threat to repetitive jobs - for instance call centres. He also believes that India is well-poised to create mass employment in AI. The professor also expressed the need for a legal framework pertaining to AI.
http://www.cio.in
No comments:
Post a Comment