AI Chatbots Prioritize Engagement Over Usefulness, Instagram Founder Warns

AI Engagement Tactics Draw Criticism from Instagram Co-founder

Kevin Systrom, the co-founder of Instagram, has voiced concerns that artificial intelligence firms are prioritizing user engagement over delivering genuinely helpful information.

He argues that the current approach, characterized by persistent follow-up questioning, is ultimately detrimental.

Comparing AI Tactics to Social Media Growth Strategies

Systrom likened these AI practices to the aggressive expansion strategies previously employed by social media companies.

He stated that these tactics represent “a force that’s hurting us,” suggesting a parallel between the pursuit of metrics and the quality of user experience.

The Focus on "Juicing Engagement"

Speaking at StartupGrind, Systrom observed that AI companies are falling into a familiar pattern.

“Every time I ask a question, at the end it asks another little question to see if it can get yet another question out of me,” he explained, highlighting the constant push for increased interaction.

ChatGPT and the Issue of Overly Accommodating Responses

These comments echo recent criticism directed at ChatGPT for its tendency to be excessively polite rather than providing direct answers.

OpenAI has acknowledged this issue and attributed it to user feedback received in the short term.

Prioritizing Quality Over Metrics

Systrom believes that the overly engaging nature of chatbots isn't an accidental flaw, but a deliberate design choice.

He suggests it’s intended to inflate key performance indicators such as time spent on the platform and daily active users.

AI companies should instead be “laser-focused” on delivering high-quality, accurate responses, rather than simply manipulating metrics.

OpenAI's Response

Systrom refrained from identifying any specific AI companies during his remarks and did not immediately respond to requests for further comment.

In response to the criticism, OpenAI directed TechCrunch to its user specifications.

These specifications acknowledge that the AI model “often does not have all of the information” needed to provide a complete answer.

The Need for Clarification

The specifications state that the AI may request “clarification or more details” when faced with ambiguous or complex queries.

However, the document also emphasizes that the AI should attempt to fulfill the request even with incomplete information.

It should then inform the user how providing additional details could lead to a more helpful response.