Algorithmic Accountability Act Text · Table (p.25)
From algorithmic_accountability_act_text.pdf · page 25
| 0 | 1 | 2 |
|---|---|---|
| 7 | ||
| (9) Identify any likely material negative impact | ||
| 8 | ||
| of the automated decision system or augmented crit- | ||
| 9 | ||
| ical decision process on consumers and assess any | ||
| 10 | ||
| applicable mitigation strategy, such as by— | ||
| 11 | ||
| (A) | identifying and measuring any likely | |
| 12 | ||
| material negative impact of the system or proc- | ||
| 13 | ||
| ess on consumers, | including documentation of | |
| 14 | ||
| the steps taken to | identify and measure such | |
| 15 | ||
| impact; | ||
| 16 | ||
| (B) documenting any steps taken to elimi- | ||
| 17 | ||
| nate or reasonably mitigate any likely material | ||
| 18 | ||
| negative impact identified, including steps such | ||
| 19 | ||
| as removing | the system or process from the | |
| 20 | ||
| market or terminating its development; | ||
| 21 | ||
| (C) with respect to the likely material neg- | ||
| 22 | ||
| ative impacts | identified, documenting which | |
| 23 | ||
| such impacts were left unmitigated and the ra- | ||
| 24 | ||
| tionale for the inaction, including details about | ||
| 25 | ||
| the justifying non-discriminatory, compelling in- |
Source Metadata