AI-Fraid For Life

Whistleblowers have often paid a heavy price for exposing the powerful, sometimes even with their lives. Was Suchir Balaji, the OpenAI whistleblower whose suspicious death in the US has taken a controversial turn, doing something similar to what Joe was doing in the cyber-thriller ‘CTRL’ and was murdered in the process?

The digital world is murky, and so are the corporations running them, as is being believed in the case of Balaji, a key figure in critical lawsuits against OpenAI and other tech companies. He was found dead in his apartment under circumstances that have been described as “highly suspicious.”

Suchir Balaji, a former artificial intelligence researcher at OpenAI, wanted to warn people, but he was found dead in his San Francisco apartment on November 26.

In an October interview with The New York Times, Balaji shared his disillusionment after realizing that OpenAI’s practices—especially its use of internet-sourced data to train AI models—could be infringing on copyright laws.

Experts believe with regulation lagging; we won’t be able to control them because they will have thought of everything we think of a million times faster than us.

Challenges & Concerns

Speaking about the increasing concerns surrounding OpenAI, Vivek Sharma, the investment head at Estee Advisors, notes that the rapid advancement of AI technologies has led to both innovation and concerns. Handling massive personal data is a major issue, since we are already giving machines bank accounts, credit cards, email accounts, and social media accounts.

He says, “AI systems use large datasets, frequently from the internet, generating data privacy and copyright concerns. Lack of openness in data gathering and utilisation heightens these concerns.” Vivek warns that misuse — whether deliberate or accidental — increases as these models get more powerful. “Many global businesses have banned AI chatbots to protect private data. Some companies are integrating customers’ personal data with AI platforms to offer more personalised solutions, raising concerns about data integrity and protection, especially given India’s weak data protection laws,” he says.

Danger of AI

Society is already dealing with a slew of issues caused by the fast spread of AI, including misinformation, polarisation, and algorithmic prejudice. Meanwhile, tech corporations compete to create even more powerful AI systems, but research into AI safety lags far behind.

As Geoffrey Hinton and other prominent figures have pointed out, concerns around data privacy, ethical use, and safety are critical and warrant ongoing attention. “While the pace of AI evolution can be intimidating, it’s important to remember that these technologies are tools created and controlled by humans. The focus should be on ensuring they are deployed thoughtfully and aligned with societal values. Collaborative efforts between policymakers, researchers, and the tech industry are essential to create robust frameworks that mitigate risks while maximising benefits,” says Sonam Srivastava, founder at Wright Research.

Regulating AI

Without giving powerful AI systems clearly defined objectives or creating robust mechanisms to keep them in check, AI may one day evade human control. “At Be Legal, we are concerned about the rapidly growing proliferation of AI technologies, particularly by OpenAI and other similar organisations. If left unchecked, the transformative potential of this cutting-edge technology can end up posing greater threats to legal, societal, and ethical frameworks,” says Ayush Jindal, advocate, Supreme Court.

AI companies aggregating users’ personal data for training models without clearly stated consent may breach principles of consent, purpose limitation, and data minimisation. “To counter these risks, we certainly need mandatory and independent audits of AI training data sets to ensure complete compliance with the privacy laws,” says Ayush.

Money Matters

A growing number of financial institutions are leveraging the power of AI for various purposes, including customer engagement, fraud detection, and investment advice. This approach presents significant risks, particularly when the tools lack transparency and reliability. Any potential misuse can result in unauthorised access to financial records. To combat this, it is necessary to mandate transparency and disclosure of data usage from AI companies. The horizons of the Consumer Protection Act, 2019, must be expanded to take into consideration the new and nuanced risks,” says Ayush, adding, “We need to develop a comprehensive legal framework around AI and also strong regulatory oversight, mandating regular compliance and checks.”

AI systems use large datasets, frequently from the internet, generating data privacy and copyright concerns.” — Vivek Sharma, Investment Head, Estee Advisors

Balaji’s Role at OpenAI

During his four years at OpenAI, Balaji worked on many major projects, including WebGPT. He was crucial in collecting and organising the internet data required to train ChatGPT. Balaji resigned from OpenAI in August 2024, citing concerns over the use of copyrighted information for AI model training. Only three months before his death, he charged OpenAI with plagiarism and copyright violation.

The horizons of the Consumer Protection Act, 2019, must be expanded to take into consideration the new and nuanced risks.” — Ayush Jindal, advocate, SC