Regulatory Frameworks for AI and Big Data: A Short Primer
By Aishvarya, Anmol
The Telecom Regulatory Authority of India (TRAI) is looking to explore how they can incorporate the use of Artificial Intelligence (AI) and Big Data (BD) in the telecom sector to enhance the overall quality of service, spectrum management, network security and reliability.
It has released a consultation paper that deals with issues relating to the adoption of AI and BD in the telecom sector and is seeking feedback from citizens on these issues before they go ahead with making any rules or regulations on this.
Before you delve into the consultation paper, here is a recap of what these terms are and what their incorporation in the telecom sector means.
The What and Why of AI and Big Data
Simply put, Artificial Intelligence (“AI”) is the intelligence demonstrated by machines. It is an array of technologies that enables machines to act with higher levels of intelligence and imitate the human capabilities of sense, comprehension and action.
TRAI expects that if AI is adopted, it would optimize and effectively utilize the existing telecom resources to improve services and digital connectivity. It could also be used to optimize network performance through self-optimization, anomaly detection and assisted capacity planning.
As for Big Data (“BD”), it refers to technological developments related to data collection, storage, analysis and applications. It is characterized by the 4 “Vs” — Veracity, Velocity, Volume, and Variety of data.
AI and BD have together revolutionized almost every single service industry by providing effective software solutions for network infrastructure, operation, maintenance, management, and network security. It improves service quality and customer satisfaction by making effective and optimal use of resources.
TRAI predicts that leveraging AI and Big Data in a harmonious manner can enhance network security, provide stability and improve spectrum management.
Opportunities that AI and Big Data Present to the Telecom Industry
The networking industry is also seeing the growth of artificial intelligence. Here, it is used to address the difficulties brought by virtualization and cloud computing through increased network automation and efficiency. The following are some of the sector-specific benefits and possibilities that AI brings:
- Quality of Service
Artificial intelligence in network infrastructure enables telecom service providers to reliably predict marketing conditions for a new geographical location and its surroundings. AI may also help prior to deployment by offering 3-D modelling of the network environment and radio performance. Minor factors such as trees and construction that traditional planning tools may overlook may also be included in these AI models. Because these elements affect signal transmission, this level of specificity is essential for the wavelength and spectrum.
In case of digital connectivity inside buildings, AI may be utilized to construct a network representation of the building, which could analyze the quality of service at each corner of the structure. In this case, the AI may learn the network’s behaviour for that building based on data from the mobile device at each location. Based on the data, the AI can anticipate the best network equipment to put in a given region to improve the building’s connectivity. Moreover, after installing the recommended device, AI may monitor device performance on a regular basis to identify elements that might affect network performance prior to any network device failure.
2. Network Security
One of the uses of AI is spectrum management (the process of regulating the use of radio frequencies to promote efficient use) and sensing (the process of periodically monitoring a specific frequency band, aiming to identify the presence or absence of users), which is essential to achieving dynamic spectrum management in wireless communication systems and is often used to assist users in understanding the condition of channel occupancy. AI can instantly analyze millions of events and identify possible cyber threats, such as malware that takes advantage of newly discovered security flaws or risky behaviour that might lead to phishing scams or the download of malicious software.
3. Customer Centric
Artificial intelligence may incorporate data from conventional and behavioural systems, location tracking, social media monitoring, and other sources in order to completely comprehend and personalize the user experience in both the offline and online worlds.
Chatbots that are virtual assistants driven by AI can help businesses save money while letting consumers solve minor issues in their own time. In order to quickly and effectively respond to client inquiries via email, chat, and phone calls in regional languages, AI-led chat and voice assistants are used. In fact, AI might make it simpler for customers to compare different plans by taking into consideration their own consumption patterns.
4. Broadcasting Sector
AI has the potential to scan the contents of films sequence by sequence and identify things to add suitable tags. As a consequence, despite the volume, every piece of information discovered by media companies becomes readily available. AI-powered platforms give customers content suited to their specific interests, resulting in a more personalized experience. Using AI-based data analysis and natural language generation-based reporting automation technologies may create performance reports with easy-to-understand analyses, providing them with accurate insights to make educated data-driven decisions.
Potential Risks of AI
The opportunities that AI and Big Data present to the telecom and other sectors are enormous but at the same time, we must not and should not overlook the risks associated with it. The most pressing ones are data biases, data poisoning and privacy violations. If data used for building an AI model is biased against a group, the model will replicate the human bias in selecting them and learn to bias against that group. Data poisoning can be used to undermine AI-powered identification of money laundering operations or create ransomware that impairs the smooth operation of an agency. AI might also be used to identify, track and monitor individuals and the people affected by it might have no recourse.
We must not wait until after the development of AI models to mitigate risks. Instead, risk analysis should be part of the initial AI model design, including the data collection and governance processes.
Regulating AI and Big Data so far
While a solid regulatory framework to support the responsible use of AI in this sector is yet to be finalized, the same has been discussed by various bodies of the government.
To strategize AI, NITI Aayog, in June 2018, recommended that research in the field should be promoted; the workforce should be skilled and then re-skilled; guidelines should be framed for ‘responsible AI’ and to facilitate the adoption of these AI solutions.
Through July 2019, MeitY issued multiple recommendations through reports on specific themes dealing with AI. Some of these recommendations are as follows —
- Develop an enhanced AI resource platform on a national level (NAIRP).
- Collate data that is publicly shareable, along with tools, information, best practices etc., to enable large masses to participate in AI tasks.
- Research and Development in cybersecurity — tools & techniques, newer vulnerabilities, challenges, anonymization, best practices and setting up a national level Resource centre for AI in cybersecurity.
- Publicize available data for AI and establish transparent and clear controls.
- Enable supportive policies and remove bottlenecks in relevant legislation.
- Enable collaborative works between POCs and large-scale technology companies along with funding from the government.
NITI Aayog through 2019–2021 worked on the theme of Responsible AI and broadly proposed risk-based regulatory interventions to build a trusted AI ecosystem; awareness drives and capacity building for the public about responsible AI; and, aligning responsible AI principles with procurement mechanisms. Further, some target policy interventions for responsible AI were to adopt -
- A custodian for responsible AI principles who would monitor the same, update these principles and identify effective enforcement mechanisms
- Research into responsible AI from the technical, social, legal and policy lenses
- Make data and responsible AI tools and techniques accessible while developing India’s perspectives on responsible AI.
As one might expect, corporations all around the world are embracing this technology. In the past few years, countries around the world have taken several initiatives to frame strategies to guide and foster the development of AI and mitigate the risks associated with it. Countries like Canada, China, Japan, Finland, United Arab Emirates, began to formulate national and regional strategies in 2017. Thereafter, in 2018, many countries such as France, Germany, the United Kingdom and European Union also published their plans and strategies regarding the adoption of AI and related subjects.
Aishvarya Rajesh, a student at National Law University, Delhi and Anmol Singh, a student at Jamia Hamdard, are interns from Civis’ Policy Leaders programme.
