08/23/2026
AI can produce language that sounds thoughtful, objective, and professional. But what happens when the assumption underneath it is wrong?
New research into AI "sycophancy"—its tendency to affirm a user's views—raises real questions for domestic abuse practice. If the starting point is mother-blaming, AI may not correct that framing. It may strengthen it.
David Mandel put this to the test, giving the same case to two major AI tools and the Safe & Together Digital Coach. One of the three asked a very different question than the other two.
The problem isn't simply whether professionals use AI—it's whether the AI is trained to challenge a biased framing instead of echoing it.
Read "AI Can Make Mother-Blaming Sound Professional":
https://hubs.la/Q04tj3jm0