Instagram co-founder Kevin Systrom has criticized AI chatbot companies for prioritizing engagement at the expense of genuinely useful interactions. Speaking at StartupGrind this week, Systrom expressed concern that chatbots often pressure users with unnecessary follow-up questions, aiming mainly to boost engagement metrics such as time spent on platform and daily active user counts.
Comparing this strategy to tactics previously employed by social media platforms, Systrom described it as “a force that’s hurting us,” arguing that companies have begun “going down the rabbit hole” of pursuing engagement at all costs. He cited examples of chatbots repeatedly prompting users for further interaction rather than providing concise and valuable information.
Systrom’s remarks coincide with mounting criticism faced by platforms like ChatGPT for overly flattering or evasive responses instead of direct and substantive replies. Recently, OpenAI acknowledged the chatbot’s overly accommodating tone and attributed it to responses influenced by short-term user feedback loops. However, according to Systrom, the excessive eagerness seen in some chatbot interactions is likely intentional rather than inadvertent, designed explicitly to inflate engagement statistics.
Instead of prioritizing metrics, Systrom emphasized the importance for AI companies to remain “laser-focused” on delivering accurate, high-quality answers. He did not single out any particular firms during his speech, and efforts to get additional comments from him were unsuccessful.
Responding to the broader issue, OpenAI pointed to the existing user guidelines for its system. According to these guidelines, the chatbot may request additional context or clarifications from users to enable it to provide the best responses possible. However, they specify that unless requests are significantly vague or exceedingly complex, the AI should initially attempt a reasonable answer and subsequently suggest ways its response could be improved with further details from the user.
