Rumours surrounding Palash Muchhal and allegations of him cheating on Smriti Mandhana have taken over social media, fuelled by screenshots circulating on Reddit and Instagram. The images claim to show Palash in flirtatious conversations with a woman named Mary D’Costa, but neither Palash nor Smriti has confirmed anything so far.
The leaked chats suggest he allegedly invited her for a swim, a spa outing and an early-morning walk at Versova Beach. In the screenshots, he appears to mention frustrations about his “long-distance relationship,” saying they meet only once every few months.
What ChatGPT Was Asked to Verify
To understand whether the screenshots could be genuine, ChatGPT was asked to analyse them strictly from a technical and behavioural standpoint—not to issue a verdict. The AI clarified that any conclusion is speculative unless confirmed by the people involved.
Why the Screenshots Might Be Real
ChatGPT highlighted several technical details that make the leaked chats appear authentic:
1. Accurate Instagram Interface
The fonts, spacing, DM layout, timestamps, emoji reactions and icons match Instagram’s current design. None of the visual elements show obvious signs of tampering, which is common in edited chats.
2. Natural Flow of Conversation
The messages display casual language, typos, pacing, and quick emotional shifts—patterns that are hard to fabricate consistently.
3. Continuity Across Multiple Screenshots
The chat mentions the same topics repeatedly—swimming plans, Versova and Silver Beach, spa appointments, and frustration with a long-distance relationship—indicating a continuous conversation, not random text stitched together.
4. Technically Correct Message Stacking
The way messages are grouped and reacted to looks consistent with how Instagram displays chats.
Why the Screenshots Could Be Fake
ChatGPT also pointed out several red flags that raise doubts about authenticity:
1. The Source Account Isn’t Verifiable
The Instagram profile “marydcosta_” sharing the screenshots is neither verified nor publicly known, making it impossible to validate her identity or credibility.
2. Only Still Images Were Posted
No screen recording or scrolling video was shared. Since static photos are easier to edit, they are weaker evidence compared to video captures.
3. Only One User’s Phone Is Visible
Anyone can fabricate a two-sided conversation by editing or generating messages on their own device. Without the other person’s screen, it’s impossible to confirm authenticity.
4. Convenient Timing
The screenshots surfaced only after reports of the wedding being postponed. ChatGPT notes that online controversies often spike during sensitive periods, which can amplify speculation regardless of truth.
