Overview
The Partnership on AI (PAI) launched a research study last October led by Jeffrey Brown on the experiences of women and minoritized individuals in the AI field. We hosted a workshop on September 9, 2021, with key stakeholders to share our findings, and hold space for reflection about individual experiences and future directions. Participants included DEI scholars, advocates, professionals, and AI practitioners.
The Big Questions on AI and DEI
At the workshop, we explored a variety of questions about diversity, equity, and inclusion in the AI field. The following emerged as key conversation topics:
1. Are AI teams different from other tech workplaces?
Well… yes and no. Some participants acknowledged that AI teams, especially those focused on research or ethics, often deal with work that is open ended or ambiguous. One participant likened their experience in an AI team to that of being in graduate school, remarking that members sometimes need to fight for funding and recognition. AI research teams in particular can differ from other tech workplaces because of how personally invested researchers are, having worked on the same questions since their time in school with one participant noting,“for many researchers, this is their life’s work.”
Workshop participants acknowledged that AI/machine learning teams are still similar to tech workplaces in that they are cis-male dominated, with one participant noting that professionals from minority backgrounds are often the lone voices raising issues related to AI’s disproportionate impact on marginalized populations.
Another distinguishing factor of the AI field, some pointed out, is the interdisciplinary nature of many AI teams which include non-traditional tech roles like sociologists, linguists, and psychologists. This, some participants believed, made it harder for professionals without a computer science background to earn recognition.
2. How can leadership address ERG fatigue?
Many participants, informed by their own experiences establishing, leading, and contributing to Employee Resource Groups (ERGs), described a variety of reasons why engaging with ERGs can contribute to employee fatigue. Participants described being overburdened with responsibilities especially as ERGs grew bigger; feeling under-appreciated and not compensated for extra work; and finding an increasing misalignment between their own goals and the organization’s expectations. One participant reported feeling a sense of powerlessness in the face of being shut down repeatedly and becoming disillusioned with the lack of recognition.
There was, however, an acknowledgement of emerging best practices such as compensating ERG leaders, and considering DEI efforts during performance reviews. One participant went on to note that in addition to compensation, leadership must support ERG leaders by sponsoring programs for their professional development in AI. Some participants valued the crucial role that ERGs play as safe spaces for employees to turn to when they need a sounding board, or want to find future collaborators.
3. What are other DEI issues that need to be addressed?
Some other ideas that the workshop touched upon that were flagged as pressing questions for the community to consider include: 1. What are ERGs best set up to do, and not do? Are they set up to create change and who holds the power in these relationships?; 2. How can AI organizations ensure their mission and business models better align with core values of diversity, equity, justice, and inclusion?; and 3. How does de-prioritizing responsible AI or ethical AI work conflict with the well-being of diverse workers?