Welcome back to our AI and Diversity & Inclusion collection, a three part blog series that dives into the relationship between Artificial Intelligence and Diversity and Inclusion topics. This week we are wrapping up with our final post, discussing some future possibilities and predictions!
As we have discovered so far in this series, Artificial Intelligence has shown great potential to be a powerful catalyst for positive change in the realm of diversity and inclusion. In this post, we will speculate on some potential use cases, future concerns, and current developments that show that AI isn't going anywhere.
Since AI can be used in a wide variety of industries to promote D&I, this post will hone in on two main areas of interest: Human Resources and Accessibility.
Human Resources
AI is being used in the HR and People Management industries as a way to automate tasks, working in tandem with HR professionals to increase productivity. Some use cases for AI assistance that are being used today includes job description generation, resume processing, interview scheduling and general applicant management. Human Resources is also putting a special focus on the applicant onboarding experience, using AI solutions to streamline the process and automate otherwise time consuming tasks.
According to Gartner, future AI innovations could include the implementation of chatbots for job applications or interviews, a less biased hiring process, and a re-definition of expectations in the industry as a whole.
Accessibility
Any and all contributions that make our world a more accessible place for those with disabilities is highly valuable, and AI has great potential to make online spaces more inclusive for all. Some of us already have this AI technology at our disposal in our homes, with Amazon Alexa or Google Home acting as a voice-based assistant to aide with tasks, screen-read, and transmit text based information. These tools can also be built into our phones and computers, making navigating these devices easier. For those who are unable to read screens, use touchpads, and/or have vision or hearing impairments, these forms of AI assistance all help drastically increase their quality of life.
You may have also seen AI at work via Youtube automatic closed captioning, which uses speech recognition to transcribe videos and create a transcript. Though the caption services are not always 100% accurate, it flourishes when working in tandem with a human to spellcheck and proofread.
Human Resources
At the federal level, there is currently no legislation in place to govern the use of AI when it comes to employment and hiring practices. However, in May 2023, the Equal Employment Opportunity Commission (EEOC) issued a guide featuring best practices for AI usage and ways to address possible discrimination against job applicants and employees. For example, an applicant must be informed and consent when AI is being used for any kind of employment assessment or screening, as well as the reasonable accommodation allowances.
Using AI tools for assessments may lead to some applicants being excluded over others, such as screening out a candidate that produces spelling errors when typing with a screening chatbot, so these too must be monitored by a human to ensure fairness.
It's important to note that though AI can provide 'human-like' responses, it should ultimately be used as a helpful tool rather than the sole deciding factor.
Accessibility
Though AI is making great strides in making the internet more accessible for all, it does need to be monitored by a human to ensure correct information is being produced. For example, when low quality audio is used or a speaker uses filler words such as 'um' or 'uh', this can confuse the transcription software and lead to a low quality transcription. The same can be said for devices that interpret speech, with misunderstandings being unavoidable when speaking to a Google Home or Alexa.
If you've taken an image based security test such as dragging a puzzle piece into the correct spot or identifying a pattern within an image set, you've encountered image recognition software. AI also has the potential to identify images and create alternative text (also known as 'alt text') for those unable to view them, helping keep people informed of web content despite any visual impairment. However, AI has not yet mastered image recognition with 100% accuracy, leading to the possibility of mislabeling. This is why currently image recognition is used as a 'human identification' test.
Lastly, another concern is that the information regarding AI and best practices is rapidly changing. Things are moving very quickly in this space, and new information is being released all the time. What was current and valid yesterday, may have been reformed the next, which can mean that implementing AI ethically comes with an inherent need for innovation and evaluation. Data must be updated frequently, tools must be re-tested, and human feedback is crucial.
How do you envision AI contributing to the advancement of diversity, equity, and inclusion efforts perhaps 5 to 10 years from now?
Begüm: Envisioning the role of AI in the advancement of diversity, equity, and inclusion efforts over the next 5 to 10 years is both exciting and nuanced. AI holds the potential to be a powerful ally in this journey, but it also demands careful consideration to avoid unintended pitfalls. Currently, AI is already making strides in promoting diversity, equity, and inclusion. Organizations are incorporating equity-focused metrics into their AI models, ensuring that outcomes are unbiased and resources are allocated fairly. By leveraging AI, we can achieve results that may surpass human capabilities. However, it's important to acknowledge that AI's impact is a reflection of the values embedded within its design. On the other hand, is it possible to reach to a bias-free state for AI? I don't think so! It is a process of finding ways to mitigate arising bias.
To maximize AI's positive contributions, we need a collaborative approach. Developers must prioritize responsible design principles, including unbiased data selection and continuous monitoring, to build AI systems that align with equity, fairness, transparency goals. Additionally, ongoing user engagement and feedback will refine these systems further. Our collective efforts, including initiatives to foster diversity and inclusion awareness, will empower users to hold AI accountable for equitable outcomes. As we move forward, I am optimistic that a combination of heightened developer awareness, evolving design practices, and user education will shape AI into a formidable force for advancing diversity, equity, and inclusion. Together, we can ensure that AI remains a tool that amplifies our commitment to a more inclusive future.
Nabanita: 10 years ago, globalization was a buzzword and people were exchanging cultures, but we didn't have the same diversity and inclusion conversations back then because it was just conceived at that time. Now, globalization is like, on steroids! And I think that's where the inclusion culture was born and that's where we discovered these shortcomings of our society. For example, in India, I'd say there are so many different languages and it's a massive example of unity. Everybody uses different languages, but we still coexist. Globally, English is the language of business, and everybody talks in English, and so is most of the AI/NLP models or even ChatGPT. The efficiency of such technologies has been very specific to languages, and hence it wasn't as global as you'd like it to be. ChatGPT changed quite a lot and it still does a good job. I mean, it can talk in my language -- I tested it! But in terms of accessibility, for example, I can say in India, there are a massive, massive number of people for whom it's a challenge to read and understand English. So in a true sense, I think we've taken a big step towards an inclusive society. I think in the next five to ten years I will see globalization in its true form, AI will play a strong role in it and it is very exciting.
Thank you again to Nabanita and Begüm for their insight throughout this series, and don't hesitate to leave a comment here or on Diversein's Linkedin page to engage in the discussion.