“Goodwill and A.I.: A chat about workforce development for the future and modernizing a 100-year-old brand”
Article below summarizes the colloquium with Edward J. “Kingfish” Lada Jr., President and Chief Executive Officer at Goodwill, on Goodwill’s workforce development with A.I.
The summary was prepared by Mackenzie Jorgensen and Wendy Xu and edited by Kristen Scott.
Introducing Edward
Born and raised in New Orleans, Edward Lada is a first-generation college graduate. Before joining Goodwill, he spent 11 years in a minority-owned property management company. In his 5th year at Goodwill, he became the President and CEO of Goodwill of Western Missouri and Eastern Kansas, and in 2021 became the President and CEO of Goodwill Keystone Area and Goodwill Keystone Foundation. During these years as a CEO, he kept pushing an initiative to improve the way Goodwill operates with advanced technology such as artificial intelligence.
Reflecting on this endeavor, he mentioned a quote from the founder of Goodwill Helms College, “Goodwill has done a great job in empowering the poor on their job opportunities, skill training, and welfare.” Edward wants to go one step further by “moving the needle significantly around and creating intergenerational wealth mobility.” He believes advanced technology like AI, robotics, and computer vision machine learning (CVML) can be a potential equalizer of labor forces, if used in the right way.
Goodwill
Goodwill is a global social enterprise with over 100 years of history. It operates mainly through Goodwill stores and donation centers where people donate apparel and goods. Donations are then sold at affordable prices with the net of those sales going back toward the mission of Goodwill. It has nearly 4,200 stores and donation centers in North America, so 80% of Americans have at least one Goodwill store or donation center within a 10-minute drive. Each Goodwill local organization has its own board of directors and is fully autonomous.
Goodwill is a leader in ESG at local, national, and global levels. In pursuing environmental sustainability, it is estimated that 75% of the donation received in North America is diverted from landfills. Goodwill Keystone Area, where Edward is the CEO, is estimated to save 453 million gallons of water in 2021 through its recycling, reuse, and repurpose efforts while diverting over approximately 70 million lbs from landfills in central and southeastern Pennsylvania. Regarding the contribution to the social sphere, Goodwill is one of the first social enterprises in the U.S. due to its socially-focused mission [1]. It helps people with barriers to employment by providing mission-driven services, including workforce training, employment services, educational programs, and other career resources. Nationally, Goodwill serves over 2 million individuals seeking employment advice, career resources, and work opportunities, and placed 123,000 people into jobs in 2021. Goodwill operates in an autonomy business model. Each local Goodwill branch operates under a local management team who reports to the local volunteer board of directors. The Goodwill International designs standards for local Goodwill, carries out advocacy and public education. In other words, local Goodwill has the full autonomy of management in various aspects.
A. Challenges Goodwill faces
Donation quality due to fast fashion as well as a plethora of other ways people can donated or sell their used goods have impacted clothing and goods donations across the country. Markets are shifting to e-commerce; online shopping takes market shares from offline stores like Goodwill. Meanwhile, a vast influx of 3rd party resellers mines the value in Goodwill’s retail shops. Thus, Edward holds that Goodwill’s conventional business model needs refinement to maximize the intent of Goodwill’s donations. On the operations side, Goodwill still follows highly manual operation processes. Edward said, “No automation is in backrooms and Goodwill Op shops.” Thus, the Goodwill operation is ripe for disruptive innovation. Last but not least, underserved populations are being left behind by the evolution of technology, as they currently have fewer opportunities to be exposed to and get hands-on experience with evolving technology like artificial intelligence.
B. How Goodwill used technology in the Workplace
Sorting and pricing of the donated clothes in Goodwill are often subjective. Inspired by AI’s gaming-changing prospects, Edward’s team in Kansas City, in partnership with York Exponential (a robotics and AI company), developed a proof-of-concept project called the MACH II, which used a computer vision machine learning (CVML) model to categorize clothes by size, type, quality, and fabric composition. The goal was to see if Goodwill could use this technology to standardize their process and become more efficient. They also created training and workforce development around advanced technologies and showed how intuitive the technology could become regardless of people’s backgrounds, which would become the Artemis Institute.
However, “the proof of concept was a mixed bag.” Edward explained that “CVML could do the work, but the hardware limitations, the time required to collect data, and the lack of data sources and synthetic data all made it a difficult problem.” The project was also costly, requiring $500,000 in tech partnership expenses and $300,000 in additional labor.
The various challenges led to the end of this 1.5-year endeavor. At the same time, Edward had to make a tough decision about his career path and decided to move to Pennsylvania and become the CEO of Goodwill Keystone Area. There, he secured $600,000 to launch the first-of-its-kind R&D endeavor at Goodwill. One project is to create 3D-printed concrete buildings with reusable glass as an additive for the mix-3D plastic printing. Another project, partnering with a group of graduate students at Penn State University, uses CVML to identify shoes from Goodwill stock for selling online. Edward also launched a mobile learning lab for technology learning, financial literacy, and workforce development skills, such as mock interviews and career exploration in Virtual Reality.
In the long term, Goodwill is building workforce apprenticeships alongside implementing advanced technologies and business experimentation. Moreover, Edward firmly believes in the long-term goal of “workforce preparedness and readiness, which should be done with intentionality.”
C. Lessons learned
Non-profit organizations (NPOs) and Universities could and should learn from one another. Edward views “universities as not just strategic partners but also necessary partners.” NPOs have to find expertise in advanced technology with limited resources. Universities can fill that void with human capital, time, and networks. Yet, the process of partnering with NPOs can be difficult at times because of complicated bureaucracy and intellectual property policies that universities may have.
Advice to researchers:
The AI community has much to learn from organizations like Goodwill, particularly when working with marginalized populations. To work with marginalized groups thoughtfully, Edward suggests that researchers think about what it is like to go about their day; “a lot of marginalized populations live day to day, which researchers should understand and work around…. rather than assuming they already know what people need.”
Also, the unintended consequences of AI technologies sometimes take time to see. The ethical side and dual-use potential for AI must be taken seriously. Edward provides an essential reminder: “If the ethical side of what you’re developing is not as important as the creativity side, then you’re already doing [AI development] wrong.” Edward explains that awareness of the ethical side of AI technology design is crucial because there is always a risk that technology could be used in a way that the designer or the involved parties did not intend. He asks that the AI community puts “an equal weight of effort in the ethical side of what you’re creating and whom you partner with, and why you want to partner in that space.” He also gives critical advice: “Don’t fall into the trap of giving false hope;” it is better to under-promise and overdeliver!
Edward also calls for data integrity. When showing a value proposal to a partner, researchers should showcase examples of “good” or “bad” data (e.g. unethically sourced, poorly documented or with many missing attributes, etc), what these entail, and suggest improved data collection and analysis methods. Regarding AI, awareness is one of the biggest barriers for the community Edward serves, including the opportunities of AI, its unintended consequences, and how it is already influencing their daily lives.
Please find the entire colloquium talk in this link:
[1] Social enterprise is a loosely described term referring to organizations applying some business practices and management tools to drive social change in a sustainable way. It includes non-profit organizations providing services and products to the market, co-operative business, and social business. (Sources: the Social Enterprise Initiative at Harvard Business School, and Western Economic Diversification via https://www.centreforsocialenterprise.com/.
