Is AI Being Used To Hurt Indian Aviation Over AI171 Crash?

The unfortunate tragedy involving Air India flight AI171 not only shocked India and the world, but will also sadly go down in history for many firsts. It was the first hull loss of a Boeing 787 Dreamliner family aircraft, which in 14 years have carried over a billion passengers—a record for shortest time to a billion passengers for widebody aircraft. The horrific tragedy was also Air India’s first widebody accident in nearly 50 years after a Boeing 747-200 crashed into the sea after lifting off from Mumbai in 1978.

The AI171 incident is also notable for how AI has been used to discuss the crash. From experienced aviation experts to armchair commentators, it’s clear from the videos on social media that many have used AI tools to create their videos, enhance images and videos of the crash or the aftermath. While real experts have interpreted results from AI tools with caution, the ones chasing views have had no qualms about pronouncing judgments on what caused the tragedy. Some others have used AI to generate fake preliminary investigation reports which have fooled many because they seem authentic—down to the technical jargon and the staid style of writing an official report.

Since air crashes are rare, most people don’t realise that investigations can take months of meticulous and painstaking effort by teams of highly experienced specialists, and collaboration between multiple agencies—sometimes across borders—before a preliminary report is released. It is normal for a final report to take a year or two. The scamsters creating and circulating these fake AI-generated ‘reports’ are gleefully making use of this lack of awareness to push their poison. We also cannot discount the possibility that state actors could be behind these nefarious efforts too—after all, damaging Indian aviation by spreading misinformation has the potential to hurt the Indian economy and the nation.

The government has been quick to respond, with the Press Information Bureau debunking these ‘reports', but they continue to spread because the interest in the crash is naturally huge and everyone wants to know what really happened. But, in the AI-era it helps to be sceptical on some aspects, and if you receive any such report on WhatsApp, check a reputed news source like the NDTV Profit because the moment the real report is released the NDTV network will be covering it. Incidentally, if anti-India state actors are involved, it may well be that humans generated some of these fake reports. Because sometimes we can attribute content to AI and think highly of AI when the work is actually done by humans.

Strange, but true. One of our recent pieces focussed on this issue. builder.AI claimed that its AI tools were creating apps on-the-fly based on user inputs when the reality was that a team of 700 engineers were manually coding what was being presented as AI creations. The authors of our piece focus on the intersection of fraud investigations and AI. While AI is transformative and can help discover patterns and trends that can aid investigators, it can also become a smokescreen for deception as with builder.AI. How can we better investigate fraud involving AI?

Here are some of the other key articles around AI from the last couple of days:

Mark Zuckerberg Debuts Meta ‘Superintelligence’ Group In Race With OpenAI, Google

Publishers Face Collapse As AI Kills Traffic And Ad Revenue, Cloudflare CEO Prince Warns

ChatGPT's Enterprise Success Against Copilot Fuels OpenAI And Microsoft's Rivalry

Apple Executives Have Held Internal Talks About Buying AI Startup Perplexity

Meta In Talks To Buy AI Voice Startup PlayAI In Push For Talent

See you again next week. And if you get any more preliminary AI171 investigation ‘reports’ on WhatsApp, then ignore it. We all want to know what went wrong with AI171, but air crash investigations take time, just like a solid AI strategy for your organisation cannot be built overnight.

-Ivor Soans