It is true that advances in data protection have been made in the United States thanks to the passage of various laws, such as the California Consumer Privacy Act (CCPA), and non-binding documents, such as the Blueprint for an AI Bill of Rights. However, there is currently no standard regulation dictating how tech companies should mitigate AI bias and discrimination.
As a result, many companies are falling behind in creating ethical tools that prioritize privacy. Nearly 80% of data scientists in the US are male and 66% are white, showing an inherent lack of diversity and demographic representation in the development of automated decision-making tools, which often leads to biased data results.
Significant improvements in design review processes are needed to ensure that technology companies consider all people when creating and modifying your products. Otherwise, organizations can risk losing customers to the competition, tarnishing their reputations, and risking serious lawsuits. According to IBM, around 85% of IT professionals believe that consumers choose companies that are transparent about how their AI algorithms are created, managed and used. We can expect that this figure will increase as more users continue to position themselves against the technology damaging and biased.
So, what should companies take into account when time to analyze your prototypes? Here are four questions development teams should ask themselves:
Have we ruled out all types of bias in our prototype?
To create effective and bias-free technology, AI teams should come up with a list of questions to ask during the review process that can help them to identify possible problems in their models.
There are many methodologies that AI teams can use to evaluate their models, but before doing so, it is essential to evaluate the ultimate goal and whether there is any group that may be disproportionately affected by the results of use of the AI.
For example, AI teams should keep in mind that the use of facial recognition technologies can inadvertently discriminate against people of color, something that happens all too often in AI algorithms. An investigation carried out by the American Civil Liberties Union in 2018 showed that the facial recognition Amazon mismatched 28 members of the United States Congress with mug shot photos. A staggering 40% of incorrect matches were people of color, even though they only make up 20% of Congress.
By asking tough questions, AI teams can find new ways to improve their models and work to prevent these situations from occurring. For example, a close examination may help them determine if they need analyze more data or if they will need a third party, such as a privacy expert, to review their product.
Plot4AI is a great resource for those who want to get started.
Have we appointed a privacy professional or advocate?
Due to the nature of their work, privacy professionals have traditionally been seen as barriers to innovation, especially when they have to review every product, document, and procedure. Rather than view a privacy department as a hindrance, organizations should view it as a critical enabler of innovation.
The companies should give priority to hiring from privacy experts and incorporate them into the design review process so you can ensure your products work for everyone, including underserved populations, securely, compliantly, and free from bias.
While the process for onboarding privacy professionals will vary depending on the nature and scope of the organization, there are some key ways to ensure the privacy team has a seat at the table. Companies should start by establishing a simple set of procedures to identify any new or changed personal data processing activities.
La key to success One of these procedures is to socialize the process with executives, as well as product managers and engineers, and ensure that they are aligned with the organization's definition of personal information. For example, while many organizations generally accept IP addresses and mobile device identifiers as personal information, outdated standards and standards may categorize them as "anonymous." Companies must be clear about what types of information are considered personal information.
In addition, organizations may believe that personal information used in their products and services poses the highest risk and should be the focus of reviews, but they should be aware that other departments, such as human resources and marketing, also process large amounts of personal information.
If an organization does not have the necessary bandwidth to hire a privacy professional for each department, they should consider appointing a privacy advocate who can detect issues and refer them to the privacy team if necessary.
Is our personnel and culture department involved?
Privacy teams should not be solely responsible for privacy in an organization. All employees who have access to personal information or who influence the treatment of it are responsible.
Broadening your hiring efforts to include candidates from different demographics and regions can bring diverse voices and perspectives. Hiring diverse employees should not be limited to entry-level and mid-level positions. A A diverse management team and board of directors are absolutely essential to represent those who cannot get into the company.
Company-wide training programs on ethics, privacy, and AI can further support an inclusive culture while raising awareness of the importance of diversity, equity, and inclusion (DEI) efforts. Only 32% of organizations require some form of DEI training for their employees, highlighting the need for improvement in this area.
Does our prototype agree with the Blueprint for an AI Bill of Rights?
In October 2022, the Biden administration published a Plan for an AI Bill of Rights, outlining key principles, with detailed steps and recommendations for developing responsible AI and minimizing discrimination in algorithms.
The guidelines include five safeguards:
- Safe and effective systems.
- algorithmic discrimination.
- Data privacy.
- Notice and explanation.
- Human alternatives, consideration and alternatives.
Although the AI Bill of Rights does not impose any specific metrics or regulations around AI, the organizations should consider it as a baseline for your own development practices. The framework can be used as a strategic resource for companies that want to learn more about ethical AI, mitigate bias, and give consumers control over their data.
The path to privacy-first AI
Technology has the potential to revolutionize society as we know it, but it will fail if it does not benefit everyone equally. As AI teams build new products or modify their current tools, it's critical that they take the necessary steps and ask the right questions to make sure they've ruled out all kinds of bias.
Creating ethical tools that prioritize privacy will always be a work in progress, but the above considerations can help companies take steps toward correct address.
