AI Revolution is Now Here with Major Ramifications for Staffing, AAP Development, and Talent Acquisition

AI Proponents Now Changing their Tune from “AI will create jobs” to “There will be a loss of jobs” …Reskilling and Upskilling Are the New Names of the Game

A new Cisco Systems-led Consortium Report (titled “The Transformational Opportunity of AI on ICT Jobs”) buttresses conversations I have been having over the last month with Silicon Valley venture capitalists and my neighbors running tech companies heavily engaged in the Artificial Intelligence (“AI”) revolution. Those conversations have been about the growing willingness of tech companies to fess up to coming AI-driven job dislocations.

While reports vary as to how many technology and communication jobs (see our BLS Employment Situation Report story, below) have been lost due to the introduction of AI technology, over 100,000 technology jobs have been lost thus far in 2024 with more promised for the back half of the year. Some analysts believe that as much as half of that loss is attributable to the deployment of AI technology. However, AI-induced job loss measurements are not well developed just yet.

Go back and read again, slowly, the title of the Report: It is not about the coming loss of jobs or about layoffs, rather AI is now a “transformational opportunity” as to “Information and Communications Technology” (“ICT”) jobs, whether in tech companies or in other companies with technology employees. One also must think about what other job titles outside of ICT workers (not addressed in the Cisco Consortium Report) will be affected by the replacement of their skills with AI software. Kudos to the Cisco Public Relations manager who wrote that Report title.

Cisco Systems, Inc. led the consortium of nine companies participating in the study of the impact of AI that also included Accenture, Cisco, eightfold ai, Google, IBM, Indeed, Intel Corp, Microsoft, and SAP. The Report was designed to address the quickly changing technology marketplace now that first-generation AI tools are being mainstreamed into corporate use. (Many people in Silicon Valley are earnest when they predict the introduction of AI technology will be as impactful today as the Industrial Revolution was when it began in the 1760s.)  The Report and those conversations yield these headlines:

• REPORT: Pervasive Impact Is Occurring: 92% of ICT (Information and Communication Technology) workers are being (not “will be”) impacted by AI tools in the marketplace today, not tomorrow.
• The Report divided ICT job roles into seven “Job Families” involving 47 specific ICT jobs and graded the impact of AI on each job into one of these three categories:
  • “High: more than 70% of the principal skills affected by AI;
  • Moderate: between 50% and 70% of the principal skills affected by AI; [and]
  • Low: Less than 50% of the principal skills affected by AI.”
• The Report went on to, among other things, examine each of the 47 jobs by describing their critical skills and identifying which skills had been overtaken by AI tools currently in the marketplace
• REPORT and CONVERSATIONS: The old AI proponent spiel that the launch of AI tools would certainly create a loss of jobs, but those losses would be offset by the creation of other newly created jobs, has morphed into the more candid reality: “There will be a loss of jobs.” Period.
• REPORT and CONVERSATIONS: “Reskilling,” “Upskilling,” and continuous learning for ICT workers is now imperative.
• REPORT: “Every ICT job becomes an AI influenced job. The integration of AI into ICT jobs marks a profound shift, promising substantial gains in efficiency, shift in skillsets for current job roles, the creation of new job roles, and notable technological advancements.”
• REPORT and CONVERSATIONS: Members of Boards of Directors, CEOs, and Chief Technology Officers now need to immediately assess every ICT employee’s skills to make a reskill, upskill, or outplacement decision, and create a Transformation Plan. Failure to do so will leave companies still deploying pre-AI software tools at a financially competitive disadvantage.
• REPORT: “The Transformational Opportunity of AI on ICT Jobs Workers in entry and mid-level job roles have the most opportunities to upskill. In our analysis of career levels, we have observed that entry level and mid-level ICT workers are at the forefront of this transformation as 37% of entry level job roles and 40% of mid-level job roles analyzed, are expected to have high levels of transformation due to AI advancement.”
• REPORT: The Cisco Consortium also conducted a detailed examination of job roles at the functional level, in seven “job families,” based on their specific technical and functional requirements: (i) Business and Management, (ii) Cybersecurity, (iii) Data Science, (iv) Design and User experience, (v) Infrastructure and Operations, (vi) Software Development, and (vii) Testing and Quality Assurance.
• REPORT: Among the seven job family groups the Cisco Consortium analyzed, the “Business and Management” family stood out with “62.5% of job roles classified as high transformation and 37.5% as moderate transformation due to AI integration. This transformation is especially anticipated in entry level job roles such as Business Analyst, Business Intelligence Analyst, Business Systems Analyst, and Customer Service Representative [internal Table reference omitted].”
• REPORT: “Top 10 technical skills expected to become less relevant:
  • Basic programming and languages
  • Content creation
  • Data management
  • Retrieval augmented generation
  • TensorFlow Natural language processing
  • Research information
  • SQL Documentation maintenance
  • Manual XML handling
  • Manual Perl scripting Integration software
  • Manual malware analysis”
• CONVERSATIONS: AI-based software has already led to a sudden marketplace glut of software development engineers. As the Cisco Consortium Report notes: “Software development lifecycle management has notably reduced time-to-market through precise planning and predictable workflows.” AI tools for code generation have eliminated the need for (not reduced) System Level I and System Level II Software Development Engineers. System Level II software developers need to be upskilled to take on work the AI tools cannot yet perform or be transferred or outplaced to non-ICT jobs as unnecessary because they have been overtaken by technology.
• CONVERSATIONS: Some job titles (within and outside of ICT jobs) will meet a fate needing only heaving thinning (like “help desk” techs, recruiters, legal researchers, marketing writers, product description writers, librarians, accountants, and bookkeepers). Other jobs – like System Engineers Level I, System Engineers Level II, paralegals, intake clerks, research clerks, diagnostic and repair engineers, monitoring and reporting engineers, and car and truck drivers – will soon vanish almost entirely—as Blacksmiths did in the U.S. after 1900 as manufacturing and catalog ordering created transformative opportunities for them.
• NOTE: The Google self-driving car has now logged over 20 million miles and has a better driving safety record than human drivers. (The Google car does not drink alcohol, text while driving, drive over the speed limit, or weave through constricting freeway traffic at excessive speeds).

What Is The Federal Government Doing to Deal with the Coming Transformational Opportunities AI Will Afford Us?

Congress has seen this day coming and has been wringing its hands calling for contingency plans to deal with the coming work dislocations, but without results thus far. Congress knows what we are facing.

At the same time, after two years of internal division within the White House over the issue, President Biden took a leading position on AI by issuing an Executive Order to all federal Executive Branch agencies to drive them to hire AI Czars and take the plunge to adopt AI for use in their agency work. See our WIR story about that Executive Order here. The Order set out an impressive and comprehensive AI plan for the federal government and the country to keep it competitive with other nations now quickly adopting AI technology.

Responding to the President’s Executive Order, just last week in fact, the federal Government Services Agency (“GSA”) sponsored a Hackathon to reimagine what federal Government websites could look like if created and managed by AI tools.

Are Recruiters an Endangered Species?

Google thought four years ago that it would build software to replace many of the recruiters in the world of digitized countries. Google, which views itself as a search company, looks at recruiting as just another kind of search need.

Today, Google is touting an AI-driven product known as Gemini for Google Workspace. The company, which was seeking, not too long ago and for many years, to hire 3,000 employees per month, claims its Gemini tool is now embedded into every stage of typical corporate recruitment and onboarding cycles. (I cannot get ads for Gemini to stop popping up in my online searches for anything I am hunting, whether information or products to buy. I think Gemini must know I work in the employment space and has fastened on to me like a starfish to a rock.)

The Gemini tool purports to match candidates with jobs and shortens onboarding time. Tracey Arnish, head of HR for Google Cloud, told Fortune magazine: “We’re constantly thinking about organization design for Google at large in recruiting. AI is at the center of every discussion that we’re having and it’s really changing the way we’re thinking about traditional HR topics for good.”

Indeed, one of the Cisco Consortium companies, has been advertising for a year that it has AI-powered recruitment search tools now available (called an AI Job Description Generator) which the company claims “leverages job data to identify relevant skills, abilities, educational credentials and job duties.” The company’s radio ads assure listeners that it can source and curate candidates targeted to the company’s needs without the need for an intervening recruiter. Indeed believes its tool is the recruiter and all that is missing is the hiring manager’s decision to hire/not hire.

Any number of media outlets over the past year have also reported dramatic losses of recruiters from the workforce, including the Wall Street Journal. The Journal reported that tech companies alone had collectively laid off 50% of their recruiters. It is hard to imagine, but a new reality is quickly emerging in front of our eyes. This is a very difficult period of transition for recruiters.

What Is a Company to Do When It Experiences Technology-Driven Job Dislocations?

Silicon Valley Tech company senior managers wake up every morning wondering what labor-saving devices or tools they can design, build, and deploy to digitize work. I remember like it was yesterday in 1987 in Silicon Valley when many of the telecommunications sector companies realized that their various “Voice Mail” and commercial-grade telephone answering machines had put tens of thousands of telephone operators, secretaries, and receptionists out of work across the country over only a 5-6 month period of time (the length of time it took to sell-in their products, install them, and turn them on at a customer site.)

My then large law firm realized within an hour the morning that we turned on our voice mail tool for the first time that we had just eliminated the need for over 50 receptionists assigned to the many floors of our many office buildings around the country. Receptionists were there to ostensibly “guard the doors” to the law firm and meet and greet occasional clients and vendors. Their primary task, however, we learned that morning, was to answer the phones and route inbound telephone calls to the proper employee while they waited to greet the occasional visitors and vendors. Within a few months, we realized that “guarding the doors” was not enough reason to leave the door guards in place any longer, once stripped of their phone answering duties (which occupied 95% of their time we discovered that morning).

As a result, we then caused all visitors to enter through a single first-floor “Visitor Center” in each of our buildings so we could then seal the reception doors on each floor, of each office tower, remove the grand reception desks, guest furniture, and capture back thousands and thousands of square feet of office space to then be put to other more productive uses (like storing paper files…then replaced not long thereafter with offices following technological advances in digital storage tools…now soon to be abandoned and returned to the landlords with the advent of remote work).

And important to the dislocation discussion, my 200 law partners and I decided in 1987/1988 that we had a duty of loyalty to our employee receptionists who were out of work through no fault of their own. So, we offered to retrain all who wished to stay into other administrative jobs (back-office jobs, researchers (the ones now being dislocated by the new AI work tools), and paralegals (now also about to be further reduced in number, if not to zero, or near zero)). And we worked with those former receptionists who wished to leave the firm by cushioning the otherwise economic impact on them by keeping them employed for a few more months while they looked for other types of work.

I, for one, felt strongly that it was our duty as the owners of the law firm to own the responsibility for all employees we dislocated through no fault of their own. Most of my law partners thought the responsibility for their unemployment did not rest only with the state government unemployment insurance systems and employment centers. And, of course, unemployment insurance in every state was a very poor “stop-gap” to cushion the economic impact of a job loss.

I have advocated ever since that day in 1987 when we plugged in the VMX tool that companies create contingency dislocation fund pools against the day the advance of technology renders some unfortunate otherwise well-performing employee “redundant” and “unnecessary.” That message alone is likely more difficult to swallow than even the loss of pay. We are pack animals. Irrelevance is a harsh reality, let alone the thought of starting over, starting something new, and something foreign. Gut-wrenching. My heartfelt concern goes out to my many recruiter friends.

My then-client Hewlett-Packard, led by David Packard (co-founder of Hewlett-Packard), was a thought leader within the United States as to Human Resources policies and created many of the foundation HR policies and practices we take for granted today. They were revolutionary in their opening days. Mr. Packard, too, was troubled by technology dislocations as he saw a “roller coaster” ride in the 1970s, 1980s, and early 1990s of “boom and bust” cycles in the semiconductor industry (then the essence of “Silicon” Valley before software emerged from the mist to create the Internet Boom which then very quickly diversified the number of industries in the Valley which today we collectively call “Tech Companies”).

While it does not fare well these days in “employee” vs “independent contractor” law, Mr. Packard developed the concept of “doughnut” staffing in response to advancing technology changes impacting the workforce. This allowed HP to hire “mirror image” independent contractors with skills parallel to those of full-time regular HP “core employees.” When technology changed or demand for semiconductors subsided, HP could adjust to changing staffing needs to protect its “core employees” by laying off first the independent contractors and thus protecting from job loss “core” HP employees.

Coming Affirmative Action Plan Changes, and AI Changes to AAP Software Development Tools

First, as AAP Developers annually “update” client AAPs for Minorities and Women (whether in “Tech Companies” or in other companies with technology employees), they will find some Job Titles dropping out of the Workforce Analysis from last year. AAP Developers may also find entirely new Job Titles popping up where the client has newly employed AI software engineers. This change will affect not only the dependent Job Group Analyses, but also the dependent Availability Analyses, and Goals Analyses. Statistical Disparity Analyses may also change, including even as to Job Titles which persist but which have perhaps thinned out suddenly making them too small for meaningful statistical analyses.

Second, AAP Developers will find they are at a disadvantage if they do not reprogram their software using AI programming principles. One AAP Developer, with already the fastest diagnostic and assembly software in the industry (created over 39,000 AAPs in three days for one of my clients) has invested many more millions of dollars this year in AI software. Result: 13,000 Statistical Disparity Analyses which take 24 hours to generate with current software will now take only 11 seconds with the company’s new AI software. The days of shrink-wrap software are over. “Toto, I’ve a feeling we’re not in Kansas anymore.”

THIS COLUMN IS MEANT TO ASSIST IN A GENERAL UNDERSTANDING OF THE CURRENT LAW AND PRACTICE RELATING TO OFCCP. IT IS NOT TO BE REGARDED AS LEGAL ADVICE. COMPANIES OR INDIVIDUALS WITH PARTICULAR QUESTIONS SHOULD SEEK ADVICE OF COUNSEL.