AI, Telecom Infra Vital For Each Other, Concerns Remain on AI-Based System's Impact: Trai Chair

Artificial Intelligence and telecom infrastructure have become vital for each other but concerns remain around working of AI-based systems as it can affect millions of users, a top official of telecom regulator Trai said on Wednesday.

While speaking at Trai–STPI Pre-Summit Event on 'AI in Telecommunications', Trai Chairman Anil Kumar Lahoti said that telecom networks are the primary carriers of AI, and AI on the other hand provides an intelligent layer of telecom.

He said telecom service providers are increasingly using AI for network planning, optimising network operations, predicting maintenance needs and enhancing consumer experience to combat fraud, detect spam and automate complex workflows.

"At the same time, we must acknowledge that AI-based systems raise certain concerns. Automated decisions based on AI systems may impact millions of users. Therefore, making transparency, accountability and human oversight indispensable when we deal with the use of AI," Lahoti said.

He said the AI systems deployed by telecom service providers are detecting with AI and also blocking with blockchain technology a total of nearly 400 million voice calls or messages per day.

Lahoti said that with the roll-out of 5G, rapid growth in data consumption, expansion of IoT, early developments of 6G, telecom networks have become highly complex and dynamic systems. Managing this scale and complexity through conventional approaches is becoming increasingly a Herculean task.

"Artificial intelligence has therefore emerged not merely as an efficiency tool but as a foundational capability also. Artificial intelligence is therefore reshaping the telecommunication sector and especially in India where we have more than 1.2 billion telecom subscribers and around 1 billion data users," Lahoti said.

Cellular Operators Association of India (COAI) Director General S P Kochhar said that AI needs to be developed at a scale where it can benefit masses especially people in rural area with voice-based services.

"We have to build trust. Only then the benefits of AI will flow and be acceptable to the masses. Managing discussions on data ownership, which were earlier on the marginal sidelines, will now come on centre stage. Discussions on data ownership, privacy, bias, transparency, accountability, now move to centre stage," Kochhar said.

He said that pricing of AI technology is also a concern, along with risk related to the privacy and security of people.

"TRAI, your consultative and principle-based approach strikes a right balance between innovation and safeguard. We need adaptive frameworks and those frameworks must focus on transparency, security, privacy, protection, and accountability," Kochhar said.

He said there has been only discussion around these frameworks, but to implement them on ground will need a lot of brainstorming.

"We require clarity and consistency in policies and regulations, and that is what will attract investments and innovation in India. If it is not accompanied by clarity and consistency, both in policies and regulation, it will not attract the intervention of investors. That is extremely essential, and that is where we have to work overtime," Kochhar said.

He said there is a need to invest in skilled manpower that is employable.

"Now while our numbers may show we are number one in skilled manpower, we really are not. I am sorry, I am making this bold statement. I have spent six years in the skilled ecosystem. The type of people we are skilling are blue-collar workers, and they are also not in consonance with the industry requirements. There has to be a revamp of the system. We have to skill people, we have to re-skill them, and we have to up-skill them," Kochhar said.

IndiaAI COO Kavita Bhatia said that there are working groups that are taking inputs from pre-IndiaAI Impact Summit events, which will form the basis of the declaration document for the IndiaAI Impact Summit scheduled to be held in Delhi on February 19-20.

"We already have 25 working group meetings which we have concluded yesterday, the pre-summit events inputs will also be taken up by the committees and they will finalize their declaration document which will be leading up to the main declaration document," she said.

Bhatia said that there was a target to have over 300 pre-AI impact summit events for wider public participation, and till date the number of such events have already crossed 350 and 157 more pre-summit events will be held till February 7.

She said that almost 2.2 lakh people have already participated in 350 pre-summit events.

"We are expecting 15 plus country heads, 100 plus countries' participation (in the IndiaAI Impact Summit). As of today, we have confirmation from 12 countries, and we are hoping around 7 to 8 more heads of countries will participate in the summit. 50 plus global leaders will be participating (in the summit)," Bhatia said.