Will diversity and inclusion technologies beat bias?
How will diversity and inclusion technologies make a difference?
There are claims that diversity and inclusion technologies will eliminate bias and end discrimination in our workplaces. But is that really the case?
Our existing recruiting models mean that the right talent and the right job are frequently mis-matched. If they are aligned, it can be by sheer coincidence. We quite frequently hear that AI and diversity and inclusion technologies will be the panacea to all our concerns and technology will eliminate bias and discrimination in our workplaces, including talent management. It’s a nice thought, often advanced by technology vendors, although it’s not strictly true. The moment there is any human interaction inevitably bias, judgement and interpretation all come into play.
What AI and diversity and inclusion technologies can do well, is manage and monitor our systems and interactions to detect and flag up signs of bias and discrimination. This can lead to more reliable people-focused decision-making processes factoring in data, with in theory at least, reduced levels of bias.
The downside is that the very same technology may be embedded with bias due to the data sets on which the algorithms are set up, as well as the lack of diversity within the group of developers designing the various platforms.
Diversity and Inclusion technologies offer a way to monitor and manage our systems to give employees access to neutral information. This can increase awareness around unconscious bias and allow them to focus on appropriate data-based solutions. We need to be careful that when we rely on technology to do the heavy lifting, it shouldn’t mean that we stop being aware of our own biases. Systems are only a tool and it is still people who running them. Tech shouldn’t foster a culture where we assume that diversity and inclusion technologies will automatically solve all our organisational problems and we come to depend on them without any other analysis. End users have to be aware of the potential impact of the information these technologies highlight, especially if there are legal implications. Tech solutions won’t follow-up and create a get out of jail card.
At the extreme end it can fuel a big-brother mentality of resentment against excessive political correctness followed by a sense of dis-empowerment and loss of control
What Diversity and Inclusion Technologies can do
Diversity and Inclusion technologies can be helpful in a whole range of workplace situations especially in HR processes. Each piece of technology can offer a raft of benefits, but frequently come with some downsides. Without human awareness these technologies risk not being used effectively.
1. Flag up biased language
Bias kicks in right from the beginning of any selection process so diversity and inclusion technologies can be helpful to monitor employer branding documentation including web sites, job profiles and adverts. Many HR professionals are astonished when they audit their systems to find out how particularly male coded their branding is. They might toss in a few pictures of women and people from other non-dominant groups in the hope that this conveys a more inclusive image.
AI now has the ability to scan documents and communication for either inappropriate words and phrases and flag or stereotyping or biased language in all written content including job descriptions and profiles. Platforms such as Textio analyse the language in job postings and other documents to highlight gender-specific wording.
Kat Matfield’s Gender Decoder is a basic free resource which highlights male and female coded words. Although platforms such as this flag phrases such as “man-up” “grow a pair” and “ninja”, there are downsides as some basic words such as leader and acumen are still considered to be masculine coded. This is because of the number of men occupying these roles they are perceived to be male activities. If you ask anyone male or female to draw a picture of a leader, they are still likely to produce a picture of a man.
2. Talent Acquisition
We know that talent acquisition processes are riddled with bias and traditional recruitment systems relying on unstructured interviews and network referrals in particular are considered to be highly ineffective with many experts arguing that they should not be used to make final selection decisions. Yet in the US for example 65% of companies use traditional approaches to filling vacancies. Bias is reduced when selection processes reduce non-specific feedback (the infamous gut instincts we hear so much about) to give data based results. This requires trained and effective hiring teams.
Diversity and Inclusion technologies can be used in many of the key areas of the process. Here we take a look at where with a some insights into the potential downsides.
o Candidate Sourcing
A number of biases become evident at this stage of the process. Who hasn’t seen a profile with education gained in top-tier universities and employment in Fortune 500 companies and thought – great. Many companies also offer incentives from network referrals from existing employees or build relationships with the same business schools or academic establishments. But what about candidates we don’t know about?
Platforms such as Wonderkind target passive candidates by focusing on the social media and internet activity of the 80% of people who are not actively looking for jobs. Other platforms such as Entelo and HiringSolved, facilitate searches a specific set of skills or attributes. This allows organisations to “hide and anonymize candidate information commonly associated with different forms of bias before executing a search or sharing a review queue with a hiring manager.”
The downside of these approaches is that tech can still miss potential candidates who are not active on social media or who don’t pad their activities with searchable professional key words and therefore won’t appear in searches. Research from LinkedIn suggests that women tend to undersell themselves on their platform and have a higher incidence of incomplete profiles compared to male colleagues. Many older candidates are also not as active on social media as younger demographics.
Check our corporate programs to create bias conscious cultures
o Candidate Selection
The range of options in this area is significant and growing. There are a number of possibilities to target a broader talent pool and for processes to be more inclusive.
- Blind resumes/profiles: Some vendors offer the possibility to parse the information on CVs that triggers bias. This can be via removing names, gender, dates of birth, nationality even what appears to be benign information such as schools attended or even postcode. How many times have you seen Harvard, MIT, Oxford or INSEAD and thought – great background? It definitely happens. Other technologies also extract specific information and skills which might be relevant for a job application.
- Automated interviews: A number of vendors offer technology around automated interviews with standard questions which a candidate responds via an online platform, without any human interaction. Some users say this favours the Periscope generation who are practised at performances to camera and those with strong presentation skills and/or alpha personalities. Others such as Mya use chatbots to engage, screen and interview candidates. Most of the jobseekers who are interacting with chatbots today have no idea their information is being recorded as though they applied for the same position in another way. Most vendors don’t flag that up which in Europe presents GDPR issues. At some point automated interviews require a human viewing when all potential biases kick in. Another downside is a generational gap with many older workers not being comfortable in this process. Some clients have also reported gender biased user instructions with female candidates being told not to wear make-up. There is much conflicting research around women and make-up, but some suggests that women who do are considered to be more likeable and competent than those that don’t. But the Goldilocks dilemma adds yet another layer of bias in that the amount of make-up has to be just right!
- Structured interviews: These attempt to reduce the biases that occur during the interview processes by adding more structure through standardized questions and pre-determined criteria for benchmarking competences. Rather than relying on ad hoc questions, where the bias of an interviewer can be imbued into both the question itself and also how she receives the answer, it is believed that interviews should be set up so that all candidates are asked questions, in the same order and responses noted down at the time. There is usually a half way point where an analysis of the candidates performance is assessed. Interviewers are also held accountable by their colleagues for any perceptions and required to defend them. The objections to structured interviews are that the communication flow is less organic and possibly stilted, but the results are likely to be more neutral. Responses can then be compared systematically
- Facial profiling: Another dimension is facial profiling which can be done via video-interviewing. Software such as HireVue which examines facial expression for enthusiasm and soft skills and then ranks the candidates against the job profile. There are some potential downsides to this approach as well. It doesn’t factor in the mood or well-being of the candidate which might have nothing to do with their fit for the job. There could be issues in their lives which cause them to exhibit signs of unease, self-consciousness, anger or just nerves. Perhaps they are putting on a brave face if they feel unwell. It also doesn’t touch on physical disability, congenital deformity or even Botox injections. This is where human interaction would be useful. I’ve had facial profiling and it is eerily accurate.
- Nudges: Other vendors, such as Greenhouse Inclusion, use “nudges” which are “ timely reminders at a key point in the hiring process to remember to engage in efforts to promote inclusion” to reduce the potential for bias.
- Profiling of interviewers: Other platforms such as one offered by 8 and Above, integrate online video AI to assess potential biases of the interviewers by analysing speech and body language during interviews with candidates. This allows supervisors to give constructive feedback.
Download the 3Plus eBook 12 key steps to attract, recruit and retain female talent
3. Background checks
Another approach is to reduce bias in the background check process. Checkr provides features that standardize the criteria on which people are screened out and can filter out records that are not relevant for a specific role (e.g., a customer can choose for minor drug offenses not to be reported for a customer support role). GoodHire carry our relatively affordable background checks and give candidates an opportunity to explain any items which may negatively their eventual selection.
4. Mentorship/Career Management
- Finding mentors: These solutions focus on providing online tools to search for mentors, to expand the reach of diverse mentees. Enhanced results can be achieved digitally for task and goal management and to monitor progress Providers in this space include SAP SuccessFactors as well as vendors such as Chronus, Glassbreakers, Insala, and Mentorloop.
- Career coaching: 3Plus offers personalized career pathing and development, executive coaching and targeted skill development for all women.
5. Performance Management
Diversity and Inclusion technologies can also play a role in reducing bias in feedback assessment. Zugata identifies disparities in language used in performance feedback to make sure managers treat all employee reviews the same, regardless of employees’ gender or race. Research published in HBR suggests that the language used to give feedback to women is either vague with negative vocabulary than that used with men. The focus also tends to be on style rather than skill.
6. Wage gaps
To increase transparency in wages, AI will crunch data to identify trends in promotions and help to prevent wage gaps ADP – Has a pay-equity analyzer that shows if your organization is paying men more than women for the same roles. Visier also does this.
7. Voice recognition
With research indicating some non-inclusive behaviours related to interruptions and certain groups dominating discussions, AI voice recognition solutions in meeting rooms, will identify who is speaking the most or speaking over and interrupting others, will allow the voices of those that have been too often been silent, ignored or interrupted.
The Stockholm-based design firm Doberman believes the first step toward erasing gender bias in the workplace is making it known. For that, it designed an app called GenderEQ that monitors and evaluates meetings based on voice recognition, then analyzes the data to show the percentage of time taken to up by male and female speakers.
GenderTimer is an app created to promotes gender awareness in social groups and workplaces. Gender Timer indicates who speaks most and dominates the air space in your meetings. It serves as a framework for discussing standard gender speaking patterns in public places.
8. Reporting coercion or harassment via chatbot
Reporting any sort of sexism and harassment is always a challenge and as we have seen many targets are reluctant to do so for fear of repercussions. AI allows anonymity for employees when they report problems related to sexism and harassment, to allow them to report incidents or concerns without fear of reprisal. Employees are “interviewed” by an AI a chatbot rather than a person. Other technologies are being developed by organisations such as AllVoices, and Callisto which deal with sexual assault reports.
The downside of this is that it could remove the human need for rapport and empathy which can be achieved with the right type of contact.
There is no doubt that diversity and inclusion technologies can help manage the rampant bias found in our talent management systems. But they won’t eliminate it totally. At some point there has to be human interaction in any organisation. We need to supplement all our tech solutions with appropriate training for all HR, hiring managers and leaders to create bias conscious cultures.
Change the way you source, attract and hire women with Executive Search and Diversity Recruitment
Found that interesting?
Learn more about our services
Make your dreams a reality with a professional evaluation of your career to date.
The evidence is in. More women in your company can deliver 35% greater financial returns. (Catalyst)
If you would like to join the live recording of these sessions please contact [email protected] for the Zoom Link
Dates for the Diary
September 9th - Podcast recording Talkpush - Discussion recruitment for inclusive workplaces
September 21st - ENGIE Gender bias in Performance Assessment online
October 26th - Banque de Luxembourg Préjugés sexistes dans le processus de recrutment.
We have Remote Learning Programs available
Check out our exciting portfolio of offerings to support your business in upskilling and competence building for your teams, to address the unprecedented challenges that women face in this new totally a digital world.
Download and listen free podcasts
How mental health and menopause are connected – and what you can do to feel better with Nicki Williams from Happy Hormones
Is “guys” as a form of greeting a gender neutral term? Dorothy Dalton responds to a poll with absolutely not. It’s male coded.
Avoid multitasking pitfalls. Multitasking isn’t what you think. It’s not focusing on more than one task at a time; it’s task switching.