OpenAI's Dilemma with Privacy Issues

2023-09-14

AI models impersonating human behavior have come at a cost for all your searches, readings, or clicks on the internet. Whether it's your Instagram photos, conversations with chatbots like Bing, or emails, they provide large tech companies with a wealth of personal data. Despite Silicon Valley's ongoing exploitation of public data, people have raised questions specifically about the secretive activities of OpenAI, a renowned company.

The California-based AI research lab released a 98-page technical report earlier this year on GPT-4, which was considered opaque or "closed" in any meaningful aspect. The company, led by Sam Altman, did not disclose any details about the architecture (including model size), hardware, dataset scale, or training methods, making it one of the most secretive releases to date.

Emily M. Bender, a linguistics professor at the University of Washington, expressed no surprise about this secrecy on Twitter. "They deliberately turn a blind eye to the most basic risk mitigation strategies, while claiming to work for the benefit of humanity," she wrote on Twitter.

For those who now find it difficult to know if their work will be replaced, this secrecy makes the situation unfair. Additionally, proving one's intellectual property and copyright in court is nearly impossible, giving OpenAI a legal but unfair advantage.

This report claiming that the company does not want to disclose details to maintain its dominant position in the market has sparked strong reactions. OpenAI is at fault, as are other companies building AI models that simulate human work, entertainment, and creativity. We have simply been trusting these companies' words, resulting in privacy chaos.

Theoretical Security, Privacy, and Safety

After some initial struggles, OpenAI has successfully avoided being expelled from Italy by offering limited control. According to EU law, even Google's writing assistant Smart Compose, which has been trained on public Gmail data, is turned off by default. Tech companies often encounter minor legal difficulties but ultimately resolve them through compensation, finding legal loopholes, or adjusting company guidelines.

In less than a decade, tech companies including Google, Apple, Meta, and Amazon have collectively received over $30 billion in fines. Isabelle da Silva, the chairwoman of the French Competition Authority, publicly stated that fines are not just part of the operating costs for tech giants. "Fines are an element of determining wrongdoing," Silva said.

A recent investigative report by Geoffrey A. Fowler of The Washington Post raised a thought-provoking question: "Which of our data is off-limits, and which isn't?" This investigative report delves into how companies like Google use your data, which you have little control over. He further points out that most answers lie in lawsuits, investigations, and some new laws. However, at the same time, large tech companies are setting their own rules.

Setting Boundaries

The debate about tech companies and their Orwellian-style mass data collection has always existed. Their refusal to abandon these practices often sparks anger and eventually leads to regulatory intervention.

Mozilla launched a campaign calling on this software giant to explain itself. The announcement stated, "If even nine privacy experts can't understand how Microsoft handles your data, what chance does an ordinary person have?" As part of this campaign, four lawyers, three privacy experts, and two activists reviewed the service agreement updates from Microsoft, which will take effect on September 30. Surprisingly, not a single expert could determine whether Microsoft plans to use your personal data.

A year ago, the Federal Trade Commission announced a plan to combat what it considers "harmful business surveillance" or "collecting, analyzing, and obtaining information from people." Since then, there has been no progress in the case.

Tech companies walk a fine line between "improving products" and stealing. On a darker note, these AI companies have been battling internal ethicists or other ethicists worldwide.

Although OpenAI was initially a nonprofit advocate, it has changed. The company's ChatGPT has given artists and authors enough reason to take this startup to court. But its secretive practices give it an advantage over other companies. Ironically, Jack Clark, OpenAI's former policy director, said in 2019 when GPT-2 was released, "It's better to talk about the dangers of AI before they arrive than to pretend not to see them."

RECENT AI NEWS

RECENT AI TOOLS