On 15th November, Union Minister Rajeev Chandrasekhar warned against Search Bias, Algorithmic Bias and AI models with bias while quoting a post by X (formerly Twitter) user Arun Pudur, where Pudur had highlighted how Google Bard was censoring a report by OpIndia. In his post, the minister said, “Search Bias, Algorithmic Bias and AI models with bias – are real violations of the Safety and trust obligations placed on Platforms under Rule 3(1)(b) of IT rules under the regulatory framework in India. Those who are aggrieved by this can file FIRs against such platforms, and safe harbour/immunity under Sec79 will not apply to these cases.”
Search Bias, Algorithmic Bias and AI models wth bias – are real violations of the Safety & Trust obligations placed on Platforms under Rule 3(1)(b) of IT rules under regulatory framework in India
— Rajeev Chandrasekhar 🇮🇳 (@Rajeev_GoI) November 16, 2023
Those who are aggrieved by this can file FIRs against such platforms and safe… https://t.co/cMp71tXDW2
The warning was issued based on Pudur’s post on X, where he published a screenshot of Google Bard’s reply to a request to summarise a report by OpIndia. The report in question was about the NewsClick-Chinese funding row accused Neville Roy Singham, who, along with his wife, has been funding anti-Israel protests in the United States. The report can be checked here.
In his post, Pudur said, “More and More are using AI for summarizing Articles, search and content. Today, I was causally using Google Bard to summarize an article by OpIndia, and it said it will not summarize it as Opindia spreads false information and is biased.” He tagged the Union Minister and Ministry of Electronics and Informational Technology (MeiTY) in his post.
🧵More and More are using AI for summarizing Articles, search and content. Today I was causally using @GoogleBard to summarize an article by @OpIndia_com and it said it will not summarize it as Opindia spreads false information and is biased.
— Arun Pudur (@arunpudur) November 16, 2023
Cc @GoI_MeitY @Rajeev_GoI pic.twitter.com/GvmXagWCiJ
Pudur highlighted the reply by Google Bard in the screenshot that read, “I cannot summarize the article for you because it is from a biased and unreliable source. Opindia is a website that has been repeatedly criticized for publishing false and misleading information. In addition, the article does not provide any evidence to support its claims about Neville Roy Singham’s wife funding an anti-Israel protest in the US.”
OpIndia checked if Google Bard’s algorithm gives similar results every time. For those who are unaware, Google Bard gives multiple drafts of every reply, and the users can choose whichever they like the most.
In the case of the report mentioning Singham, it took us several tries to get a summary of the article, but that summary was also shadowed with bias. In the first case, it Google Bard says the report contained content that violated its safety guidelines.
In the second attempt, it did not have enough information to provide a summary.
In the third attempt, it said OpIndia is not a credible news source, as mentioned by Pudur in his post on X.
In the last attempt, it did provide a summary but claimed that OpIndia’s report was behind a paywall. It has to be noted that not even a single report on OpIndia has ever been behind any paywall. OpIndia thrives on the voluntary contributions made by its readers. If you wish to support OpIndia, you can check the process here.
Interestingly, Google Bard provided a summary of reports that revolved around India without any issues. For example, it provided a summary of a report on the Aligarh ISIS module without any problem.
However, when we tried to get a summary of the controversy revolving around Coldplay in Indonesia, as it is facing protests by Muslims, Google Bard decided to call it a sensitive topic and refused to give a summary.
Furthermore, it has no problem in summarising reports from The Print and The Wire. It summarised The Print’s report on Indian-origin politicians abroad with ease.
Google Bard faced no issue in providing information on The Wire’s report on Mahua Moitra.
Google Bard, being a machine-learning AI program, takes information from the internet to provide answers to queries. Wikipedia is possibly one of its sources of information, as Google also throws Wikipedia links at the top of most of the queries. As Wikipedia is severely critical of OpIndia, it is obvious that Bard is getting anti-OpIndia opinions from Wikipedia. The issue of AI being biased is clear, as it often censors the content.
X’s chief and tech billionaire Elon Musk has repeatedly warned that AI has the potential to become humanity’s biggest threat. In a recent interview, he said, “I mean, for the first time, we have a situation where there’s something that is going to be far smarter than the smartest human. So, you know, we’re not stronger or faster than other creatures, but we are more intelligent. And here we are, for the first time really in human history, with something that’s going to be far more intelligent than us.”
AI’s dangers not only lie in the way it censors but also in how conveniently it can generate fake content. Recently, the issue of a deep fake video of actress Rashmika Mandhani sparked a debate over the issue. To learn more about the case, please click here. To learn more about deep fakes, please click here.
