Artificial Intelligence (AI) has always been viewed with suspicion by many analysists. The primary reason for this is the fact that AI lacks the empathy and reactive intelligence to think outside the parameters within which it is programmed to function. Thus, there have been many cases where AI has reproduced unacceptable forms of discrimination.
Facebook was recently indicted for reproducing gender stereotypes in the way its AI system functions in displaying adverts. It is no secret that Facebook is a global leader in the
development and deployment of exponential technologies. If they have had issues with such gender-based discrimination, then no organization is immune to such pitfalls.
Image Source: NY Post
What typically happens is that the AI is deployed with the economic motive in mind. It is often about selling as much as possible to the largest audience imaginable. Hence, the AI is programmed to think and function along such profit-driven goals and agendas. However, somehow along the line, there might be errors that would expose the inherent biases and other unpardonable issues that clearly go against certain demographic groups.
Tech companies are known to be progressive in outlook. However, minor flukes like biases and discriminatory practices might push them into a very problematic reputational position. Therefore, it is important for businesses like yours to be mindful and take reasonable care to avoid such potentially damaging reputational situations and matters.
Facebook’s Sexist Algorithms
Global Witness, a campaign group, critiqued Facebook’s job advertising system to ascertain whether there are gender biases reproduced by the system. They found that Facebook allowed customers to set gender-based restrictions - like showing adverts to men only or women only. Also, paying customers could ask for ads to be blocked from people of certain age categories. This is particularly troubling in the case of job advertisements because every major country has laws that prevent gender-biases in employment policies of companies.
Secondly, there were stereotypes in the way Facebook deployed its algorithms. For instance, some in-built algorithms meant that apps showed adverts to certain typical demographics. For instance, in adverts for nursery schoolteachers, almost all the people reached by the Facebook algorithm were women. In another job posting for mechanics, almost everyone reached was a man. This shows the possibility of AI and social media adverts from sites like Facebook susceptible in reproducing ingrained prejudices and stereotypes. And with the power of exponential technologies, they may be reproducing “exponential discrimination”.
Laws on artificial intelligence and other forms of exponential technologies are still evolving. Therefore, it will take time for a decisive legal framework to be formulated to deal with such risks. This will come with a lot of test cases and political will to control and regulate exponential technologies.
Is your Business Guilty?
Your business might have taken reasonable measures to stay compliant with all necessary employment laws including anti-discrimination. However, the use of AI and exponential technologies present novel risks and challenges that you need to be mindful of.
As we adopt new technologies, there is a strategic need to critique and evaluate risks. Any firm that fails to do this is likely to suffer a major reputational damage. This is due to the fact that technologies are supposed to be used responsibly and every major business must make a commitment to attain that end.
Your business will therefore need to undertake a critical analysis and review of risks in its new technologies to detect and restrict these risks. Online job hiring continues to grow. Therefore, positive discrimination might be the best way to stay watchful and take action quickly.
Who is Responsible?
Lawyers maintain that Facebook has committed no wrong in the way it uses its job advertising algorithms. Ultimately, the AI system serves the client and the client is responsible for their gender-based choices. This means any legal responsibility is borne by the company that uses their system in a way that directly or indirectly leads to gender-based exclusion and other forms of discrimination like racism and ageism.
In spite of this, Facebook will be amoral if they continue to allow this trend to continue. Thus, they will be compelled to find solutions to this. This is an ethical challenge and a company like Facebook is most likely to be looking into such gaps and issues.
How do you Deal with it? (Strategic Implications)
A company is responsible for their social and environmental footprints. If a company’s actions culminate in direct or indirect discrimination, they have to work to fix it. This is one of the agency responsibilities of directors. The directors of a company are required to create an ethical framework that is fair to all stakeholders.
This leads a firm like yours with the option of undertaking regular digital audits and digital risk analysis. There should be a director in charge of that (usually a non-executive director as per the principles of corporate governance). If your company is too small, it is essential to seek advice from experts who will identify risk areas where AI and exponential technologies might cause your company to act amorally. From there, you can take reasonable measures to block such risks.
Typical solutions include incremental coding and the integration of algorithms that can limit risks like discrimination in hiring and managing human resources in your firm. This might require hiring an appropriately qualified and experienced tech company to undertake the review and analysis of your IT system. This is in the domain of the strategic technology management unit of a firm, and it is always advisable to have such a unit in any major business in the 21st Century.