“Don’t do this…” Perplexity CEO Aravind Srinivas warns users after Comet browser completes online course

Aravind Srinivas

Perplexity CEO Aravind Srinivas has cautioned users against misusing the company’s AI browser, Comet, after a developer used it to automatically complete an online course.

It all started when Amrit Nigam, a web developer, posted a video on X (formerly Twitter) showing Comet AI answering questions and finishing a Coursera course titled “AI Ethics, Responsibility and Creativity.” He proudly wrote, “Just completed my Coursera course,” tagging Srinivas, Perplexity, and Comet AI.

But Srinivas was quick to step in, replying, “Absolutely don’t do this.”

His comment quickly went viral, sparking a wave of reactions online. One user wrote, “It’s either Perplexity limits this, or people will continue to do it. Telling someone ‘don’t do this’ is just shouting into the void.” Another said, “What’s the point of enrolling in a course if you don’t complete it yourself?” Others admitted that many students use Comet to get through “weekly quizzes, assignments, and mundane stuff.”

Comet AI, launched in July, is Perplexity’s new browser that aims to make web browsing smarter and more interactive. It became free for everyone in October. The browser can summarise web pages, conduct research, write emails, fill out forms, and even handle tasks like booking hotels — all through simple commands. It also remembers users’ tabs and browsing habits to offer a more personalised experience.

Comet currently runs on Windows and macOS, with a mobile version featuring voice commands on the way.

Interestingly, this isn’t the first time Comet has made headlines. Earlier, a Reddit user revealed that the AI browser executed real stock trades on Zerodha after being asked to “buy MOCAPITAL, qty: 50,” and even applied for an IPO automatically. Srinivas had then called it “one of the most exciting things about Comet,” adding, “The world of agents is here!”

However, his latest warning shows that while AI tools like Comet can be incredibly powerful, they also raise questions about ethics, accountability, and how far automation should go.