As AI becomes more and more popular in the global community, there is a growing focus on how to develop these systems.
Enter the role of an algorithm auditor, a position that entails checking the product for any potential defects or biases.
This is a task that companies must take seriously: the EU's Artificial Intelligence Act requires models with greater social impact to meet higher risk management standards.
Joe Biden, USAThe executive order on AI includes a provision requiring developers to share their red team results with ** to identify any potential risks.
This is the background to the founding of the International Association of Algorithm Auditors at the end of this year. The organization was founded by a group of leaders in the field of technology ethicsIt aims to be a central hub for auditors by providing resources, training, and eventual certification standards. But it also wants to play a role in developing audit policies that companies must consider when developing AI governance projects. Gemma Gladon Cl**ell, founder and CEO of Eticas Consulting and one of the key members of the IAAA, said the need for an association became apparent last year when she noticed that many policymakers were discussing similar topics around AI, but did not communicate these similarities with each other. She and a few other members gathered in Barcelona and Brussels to meet with these stakeholders, who encouraged them to set up a formal body that could be consulted. "One of the things we've seen is," she saidA lot of people provide companies with these AI governance platforms to integrate, and they often work in a number of different ways. None of them are subject to any form of external control, and even some of our clients will say, 'We invested in this platform and it should do bias identification.'" 
Gladon Cl**ell continued: "Therefore,How to organize and create incentives for better accounting and bookkeeping, as well as better documentation of AI processes, all of which do not yet exist. "So, the audit community, we can create these incentives across the industry to make things better, so that when we come in and check, we can verify that these things are being done well." ”Frequent internal audits are conducted within the company to ensure compliance with regulations. But said Shea Brown, founder and CEO of Babl AI and another core member of IAAAThese assessments usually have an internally focused objective designed to protect the company. In contrast, he said, algorithmic audits are externally focused. How to learnAudit, Risk & Compliance？LearnCIPM Courses, applicable to:Practitioners in risk management, privacy operations, auditing, privacy analysis, segregation of responsibilities, etc.
"Regulators, the public and everybody are asking these systems not to hurt people outside the company," he said. We care about whether this algorithm will harm people or society. ”
With the public's pairingArtificial intelligence, especially the generation system,How it can be used to spread disinformation or amplify concerns about bias against marginalized populations, there is a growing need for accountability. An international survey conducted by KPMG Consulting and the University of Queensland in Australia found that 3 in 5 respondents are wary of trusting AI systems, although 3 in 2 are also optimistic that AI systems can improve their lives.
The Pew Research Center found that there is growing concern about the use of AI in the U.S., with most people believing it will negatively impact areas such as personal privacy.
The IAAA is particularly focused on setting ethical standards for AI. It promotes the resources of the Ada Lovelace Institute and the research organization Data & Society, both of which are concerned about how to use data and AI equitably, and improve society. Both eticas and Babl AI offer AI bias audit services. These third-party audit services will be the primary focus of the IAAA, but Gladon Cl**ell said the organization would also becomeResources for freelancers and internal auditors, hoping to professionalize the industry. This means also providing an enforcement mechanism, such as an investigation of members who do not comply with established standards. She saidJust as we don't want vaccines to go through clinical trials, we don't want AI systems to go unaudited。We need to ensure that the audit process helps these systems to better benefit society. "The article is organized by the Privacy Compliance Exchange Circle, **Please note the source.