Experts have warned Australian charities that they must use artificial intelligence carefully or risk losing public trust.
Dr Kim Weinert and Adjunct Associate Professor Brydon Wang from the University of Queensland investigated the adoption and governance of AI by Australian charities and found that the sector’s gains come with certain risks.
“The philanthropic sector is under increasing pressure and, understandably, interventions that save time, effort and money are very attractive,” Dr Weinert said.
“Charities are increasingly leveraging AI for administrative, communications, fundraising and compliance tasks, as well as prioritizing where to direct their resources.
“This means that algorithms are directly involved in decisions about clients’ access to support.”
Dr. Weinert said this could be problematic.
“Charities can exercise power over people who depend on their support and do not have the ability to challenge decisions or seek help elsewhere,” she said.
“Although algorithmic standards appear to be neutral, they can indirectly discriminate against already disadvantaged individuals and groups.
“Vulnerable people and groups risk being excluded from support or being harmed by mistakes made by AI.
“Furthermore, Australian charity governance is built on fiduciary duties designed for human decision-makers, so when technology is involved in or drives key decisions instead of humans, the chain of responsibility can break down.”
Dr Wang said charities ultimately only work if they earn and maintain public trust, but trust is often confused with credibility.
“Trust is a willingness to be vulnerable to the actions of another person, whereas trustworthiness is evidenced by signals that indicate you are worthy of that trust,” Dr. Wang said.
“Good intentions and mission statements have their limits. Charities should not assume that trust already exists simply because the vulnerable communities they serve may not have viable alternatives.
“The key question when deploying AI systems is therefore not whether an organization is trusted, but whether it is behaving in a trustworthy manner.”
The researchers propose a trust-based framework developed by Dr. Wang to help charities assess whether their use of AI is justified.
Proposing a reliable AI framework
mercy – AI systems must contribute not only to organizational efficiency but also to the interests of end users and beneficiaries. This is fundamental to why charities exist.
sincerity – The use of AI must be in line with the values of the communities it serves and the aims of the charity, while meeting legal and human rights obligations.
ability – Charities should only use AI systems that they understand and oversee, and where meaningful monitoring is not possible, they may be able to demonstrate “competence” through restraint by choosing not to deploy them.
Dr. Weinert said the proposed framework will redirect decision-making in the philanthropic sector.
“While AI may bring increased productivity and efficiency, human rights, the public interest and accountability must remain at the center of all considerations regarding technology,” she said.
“We urge charities that use AI to be careful, prudent and transparent to ensure it benefits their operations and the people they support.”
Dr Weinert is from UQ’s Law School and Dr Wang is an adjunct associate professor at UQ’s Center for the Future of Policy.
The full white paper can be accessed on the Center for Policy Futures website.
