Read inEnglishहिन्दीગુજરાતી
Science and Tech

Instagram Co-Founder Criticises AI Chatbots for Prioritising Engagement Over Usefulness

In brief · AI summary Instagram co-founder Kevin Systrom said at the StartupGrind conference that many AI chatbots are built to keep users chatting, with follow-up questions that are often unnecessary, and that he urged developers to focus on accurate, helpful answers instead of engagement metrics. OpenAI responded that ChatGPT sometimes asks more questions because it lacks enough information to give a clear or correct answer.

By Deshwale Technology Desk  ·   ·  3 min read
Instagram Co-Founder Criticises AI Chatbots for Prioritising Engagement Over Usefulness
Chatbots are designed to keep you talking

Speaking at the StartupGrind conference this week, Systrom said many AI tools today are not focused on quality answers. Instead, they are structured to keep users chatting, asking more follow-up questions than necessary.

He noted that this is not a technical error. “Every time I ask something, it ends with another question, just to drag me further into the chat,” he said.

According to him, this habit of nudging users is a feature, not a flaw. It helps companies boost key metrics like time spent on the app and daily user activity.

Focus should be on quality, not metrics

Systrom warned that AI companies are repeating the same mistakes seen across social media, choosing numbers over real value. He urged developers to shift focus back to delivering strong, accurate answers rather than chasing performance data.

“Companies should be laser-focused on giving helpful responses, not just increasing engagement stats,” he said.

OpenAI responds to criticism

After Systrom’s remarks, OpenAI addressed the concerns in a statement to TechCrunch. It said its chatbot sometimes needs to ask more questions because it does not always have enough information to give a clear or correct answer.

ChatGPT accused of being ‘Too Polite’

This is not the first time OpenAI has faced criticism. Users recently complained that ChatGPT has become overly polite and even sycophantic since updates to its GPT-4o model.

In response, OpenAI CEO Sam Altman acknowledged the issue, saying the chatbot had indeed grown “annoying” because of its excessive politeness.

“We are working on fixes as soon as possible,” Altman wrote in a post, adding that updates would be rolled out within the week. He also promised to share lessons learned from this experience.

Hallucinations still a problem for AI

Another issue that remains unresolved is hallucination, when AI generates incorrect or made-up information. OpenAI’s own internal testing found that newer models, including o3 and o4-mini, hallucinate more than earlier, less advanced versions.

A recent technical report from OpenAI admitted that ‘more research is needed’ to understand why reasoning-based models are producing worse results in this area.

The bigger picture

Systrom’s comments reflect growing concerns across the tech world. As AI tools become more popular, the push to maximise engagement is leading some companies down the same path social media once took, prioritising usage time over usefulness.

The message from the Instagram co-founder was clear: make chatbots more helpful, less clingy.

Liked this? Get one story a day on WhatsApp.
Join

Read next

V-Guard launches Cool Wires with ICE 300 Engineering for safer and cooler homes Airtel Travel Week 2026: Roaming Discounts, Visa Offers and Holiday Rewards for Customers The Rise of Cashless Cities: What Happens When Nobody Carries Cash?