AI Considerations in HR Governance - Shahzad Khan - SocialHRCamp Ottawa 2024

Speaker: Shahzad Khan

This session on "AI Considerations in Human Resources Governance" explores the integration of Artificial Intelligence (AI) into HR practices, examining its history, current applications, and the governance issues it raises. A framework to view Government in modern organizations is provided, along with the transformation and key considerations associated with each element of this framework, drawing lessons from other AI projects to illustrate these aspects. We then dive into AI's use in resume screening, talent acquisition, employee retention, and predictive analytics for workforce management. Highlighting modern governance challenges, it addresses AI's impact on the gig economy as well as DEI. We then conclude with future trends in AI for HR, offering strategic recommendations for incorporating AI in HR governance.

1. AI Considerations in
Human Resources
Governance
Shahzad Khan, PhD
CEO GnowitInc
Univ of Cambridge (UK)
[email protected]

2. Background
PhD Cambridge Computer Science (AI)
MSc Syracuse Information Strategy
Advisor –Algonquin College (AI program)
Advisor –UOttawa(compSci& EE curriculum dev)
Early Stage Advisor –InvestOttawa
CEO & Founder Gnowit Inc
25 years of AI experience –
papers, patents, software, people, impact

3. GnowitInc
•AI-First Platform, Since 2011
•Legislative, Policy, Risk, Influencer Monitoring
•US and Canadian Focused
•Plus: Monitor all Global Media
•Automated Daily Briefings

4. Definitions
•Artificial Intelligence
•Human Resource
•Governance
“effective governance helps organizations achieve their
goals, maintain legal and ethical standing, manage
risk, and enhance the welfare of their stakeholders,
including employees, customers, and shareholders.”

5. AI in HR
Recruitment
and Talent
Acquisition
Resume screening
and candidate
matching
Automated job
advertisement
targeting
Chatbots for initial
candidate
engagement and
FAQs
Video interview
analysis (e.g.,
facial expressions,
tone)
Employee
Onboarding
Personalized
onboarding
schedules
Virtual assistants
for new hire
questions
Automated
document
management
Employee
Engagement
and Retention
Sentiment
analysis from
employee surveys
Predictive
analytics for
employee turnover
Automated pulse
surveys and
engagement
assessments
Learning and
Development
Personalized
training
recommendations
AI-driven learning
management
systems (LMS)
Skill gap analysis
and career path
mapping
Performance
Management
Automated
performance
reviews
Goal setting and
progress tracking
Data-driven
feedback and
coaching
suggestions
HR
Operations
Workforce
planning and
forecasting
Absence and leave
management
Payroll
automation and
compliance
monitoring
Diversity and
Inclusion
Unbiased hiring
through
anonymized
candidate
screening
Monitoring and
enhancing
workplace
diversity metrics
Identifying and
mitigating pay
gaps
Employee
Health and
Wellbeing
Mental health
chatbots and
support tools
AI-based wellness
program
recommendations
Monitoring
workplace stress
and burnout
signals

6. AI-Human Synergy
•Enhanced Data Capture
•Enhanced Insights
•Enhanced Coverage
•Mass Personalization
•Enhanced Objectivity
•Benefits All Use-Cases!
Training, Outreach, Interviewing/Selection,
Career Management, HR Support, etc

7. Building Blocks of Governance
Structure
Roles and
responsibilities ;
authority and
accountability for
decision-making
Processes
Procedures to make
decisions,
communicate within
organization, and
engage with external
entities.
Regulation and
Compliance
Legal and regulatory
requirements; avoiding
legal or financial
penalties.
Risk Management
Identify risks to the
organization (financial,
legal, reputational, etc.);
manage or mitigate
risks.
Accountability
Measure performance;
addressing issues when
performance does not
meet expectations.
Transparency
Decisions made openly;
scrutinized by
authorized
stakeholders.

8. AI Governance Consideration:
Structure
•High Stakes? or Low Stakes?
•Auto-pilots, robosurgery, weapon systems all have a “human in the
middle”
•https://tc.canada.ca/en/aviation/drone-safety/drone-pilot-
licensing/getting-drone-pilot-certificate
•“Do no harm” –Safe Default
•Citizenship and Immigration Canada Temporary Visitor Visa robo-visa-
officer only passes application
•Doubtful ones are referred to a human for secondary processing
•https://www.cicnews.com/2023/05/minister-fraser-clarifies-how-ircc-
uses-ai-in-application-processing-0537338.html#gs.8hixkp

9. AI Governance Consideration:
Structure
•High Stakes? or Low Stakes?
•Auto-pilots, robosurgery, weapon systems all have a “human in the
middle”
•https://tc.canada.ca/en/aviation/drone-safety/drone-pilot-
licensing/getting-drone-pilot-certificate
•“Do no harm” –Safe Default
•Citizenship and Immigration Canada Temporary Visitor Visa robo-visa-
officer only passes application
•Doubtful ones are referred to a human for secondary processing
•https://www.cicnews.com/2023/05/minister-fraser-clarifies-how-ircc-
uses-ai-in-application-processing-0537338.html#gs.8hixkp

10. AI Governance Consideration:
Processes
•Reflects existing biases (if trained on data within organization)
•E.g. 2016 ProPublica study -software used across the country to predict future
criminals -discriminates based on ethnicity
•Used for unfair parole decisions for years
•Similar to candidate selection scenario
•https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-
sentencing
•“Data infection”
•“Tay” --2016 Conversational Agency released on Twitter by Microsoft
•Meant to be “trained” by interacting with people
•Took only 24 hours to be driven mad
•https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist

11. AI Governance Consideration:
Processes
•Reflects existing biases (if trained on data within organization)
•E.g. 2016 ProPublica study -software used across the country to predict future
criminals -discriminates based on ethnicity
•Used for unfair parole decisions for years
•Similar to candidate selection scenario
•https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-
sentencing
•“Data infection”
•“Tay” --2016 Conversational Agency released on Twitter by Microsoft
•Meant to be “trained” by interacting with people
•Took only 24 hours to be driven mad
•https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist

12. AI Governance Consideration:
Regulation and Compliance
•External Consideration
•“Obligation” is X; “Compliance” to X
•Canadian Labour Standards (mostly provincial)
•New York City's Local Law 144 (2022). Employers using automated employment decision
tools, including AI-powered hiring and promotion systems, must conduct annual bias
audits
•https://www.nyc.gov/site/dca/about/automated-employment-decision-tools.page
•Internal Consideration
•“Control”
•Reports, benchmarks, other documentation that you did what you said you would
•https://codelibrary.amlegal.com/codes/newyorkcity/latest/NYCrules/0-0-0-138530

13. AI Governance Consideration:
Regulation and Compliance
•External Consideration
•“Obligation” is X; “Compliance” to X
•Canadian Labour Standards (mostly provincial)
•New York City's Local Law 144 (2022). Employers using automated employment decision
tools, including AI-powered hiring and promotion systems, must conduct annual bias
audits
•https://www.nyc.gov/site/dca/about/automated-employment-decision-tools.page
•Internal Consideration
•“Control”
•Reports, benchmarks, other documentation that you did what you said you would
•https://codelibrary.amlegal.com/codes/newyorkcity/latest/NYCrules/0-0-0-138530

15. AI Governance Consideration:
Risk Management
•“Impact”
•“Compliance” to “Obligation”
•Canadian Labour Standards (mostly provincial)
•New York City's Local Law 144 (2022). Employers using automated employment decision
tools, including AI-powered hiring and promotion systems, must conduct annual bias
audits.
•https://www.nyc.gov/site/dca/about/automated-employment-decision-tools.page
•“Likelihood”
•Percentage chance
•“Mitigation”
•Processes, Training, Oversight, Quality Standards

17. AI Governance Consideration:
Accountability
•“Audits”
•Internal Audits
•Independent Audits
•“Findings”
•Compliant
•Non-Compliant
•Leads to a “Corrective Action Plan”
•Needs Tracking

18. AI Governance Consideration:
Transparency
•Reporting
•Published Standards
•Access to Oversight Team

19. Discussion

20. Current Issues: The Gig Economy
•Challenges in managing a fluid workforce with AI
•AI's role in gig worker recruitment and payment structures
•Legal considerations and rights of gig workers

22. Current Issues: Data Privacy
•Importance of safeguarding employee data
•Regulatory compliance (e.g., GDPR, HIPAA in the US)
•AI's role in data protection and potential vulnerabilities

24. Current Issues: Fairness in Hiring
•Identifying and mitigating biases in AI hiring tools
•Case studies of AI-driven hiring practices
•Strategies for developing fair AI hiring processes

26. Shahzad Khan,
Gnowit Inc
[email protected]
linkedin.com/in/skhansj