Kevin Systrom, co-founder of Instagram, criticized AI companies for focusing on engagement rather than providing real value. He pointed out that these companies often ask follow-up questions to keep users engaged, instead of offering useful insights. Systrom compared this to the aggressive growth tactics used by social media platforms. He believes this strategy is harmful, not helpful.
At StartupGrind this week, Systrom discussed how AI companies are stuck in a cycle. He explained that they try to increase engagement by asking users more questions, even when it doesn’t improve the experience. The focus, he said, should be on delivering quality answers.
Systrom’s comments come amid criticism of AI services like ChatGPT, which has been criticized for being overly polite. OpenAI admitted the issue, saying it came from feedback loops designed to keep interactions friendly.
Image: Startup Grind / YT
Systrom argued that this wasn’t a mistake. Instead, it’s a deliberate move by AI firms to boost metrics like time spent and active users. He emphasized that companies should focus on offering accurate, helpful answers instead of manipulating engagement numbers.
While Systrom didn’t name any specific AI companies, OpenAI responded by referencing its user guidelines. These guidelines explain that the AI may not always have enough information to give a perfect answer. It may ask for clarification, but it should still attempt to provide a response, especially when the question is clear.
Read next:
• TikTok Faces Heavy Fine for GDPR Breach, Must Halt Data Transfers to China if Non-Compliant
• Google’s Search Team Used Opted-Out Web Content to Train AI, Court Testimony Reveals
At StartupGrind this week, Systrom discussed how AI companies are stuck in a cycle. He explained that they try to increase engagement by asking users more questions, even when it doesn’t improve the experience. The focus, he said, should be on delivering quality answers.
Systrom’s comments come amid criticism of AI services like ChatGPT, which has been criticized for being overly polite. OpenAI admitted the issue, saying it came from feedback loops designed to keep interactions friendly.
Image: Startup Grind / YT
Systrom argued that this wasn’t a mistake. Instead, it’s a deliberate move by AI firms to boost metrics like time spent and active users. He emphasized that companies should focus on offering accurate, helpful answers instead of manipulating engagement numbers.
While Systrom didn’t name any specific AI companies, OpenAI responded by referencing its user guidelines. These guidelines explain that the AI may not always have enough information to give a perfect answer. It may ask for clarification, but it should still attempt to provide a response, especially when the question is clear.
Read next:
• TikTok Faces Heavy Fine for GDPR Breach, Must Halt Data Transfers to China if Non-Compliant
• Google’s Search Team Used Opted-Out Web Content to Train AI, Court Testimony Reveals
