The news is full of how rapidly AI is evolving, with a spate of recent articles predicting mass layoffs and massive disruptions. Advances are certainly breathtaking – when OpenAI launched ChatGPT in November 2022, it added a million users in 5 days; in April 2025, it achieved this in one hour; and reportedly one-third of companies are using AI agents to automate and improve workflows.
For industrial distributors and manufacturers, the prospective shift from generative AI to Agentic AI moves deployments beyond the limited use cases of inventory management, logistics and supply chain optimization, personalized customer support, procurement and relationship management, and generating targeted marketing emails based on customer attributes to broader corporate transformational applications, across functions.
This can have far-reaching implications, as highlighted by a recent article in Axios, “Behind the Curtain: A White Collar Bloodbath,” by Dario Amodei, the CEO of Anthropic (the company behind Claude, a competitor to ChatGPT), who predicts that:
- “AI could wipe out half of all entry-level white-collar jobs — and spike unemployment to 10-20% in the next one to five years. Cancer is cured, the economy grows at 10% a year, the budget is balanced — yet 20% of people don’t have jobs.”
Amodei and others are calling for more attention to the potential short-term disruptions, especially for entry level workers who previously cut their teeth and gained experience on processes that will now become fully automated. The fear is that almost overnight, businesses will realize significant cost savings and productivity improvements from replacing humans en mass and large swaths of white-collar jobs, first in technology, then finance, law, and consulting, will be replaced by AI as the technology shifts from augmenting jobs to enhance efficiency, to Agentic AI where AI automates and does the job – instantly, nonstop, 24x7x365, and much cheaper.
Breaking the lower rungs of a career ladder can be problematic – can someone then jump mid-way up the ladder without initial experience? Could this make the trades more attractive as a career, helping backfill chronic shortages?
From another perspective, the recent pandemic/WFH trend deeply impacted the office construction segment where many white-collar workers worked, and, if more white-collar jobs disappear, that could make the pandemic collapse look mild (we will not need offices, only data centers.)
The beginning of the bloodbath could be happening now as tech companies are laying off staff to make room for AI. Recently, Microsoft laid off 6,000 across Xbox, LinkedIn, Azure units; Google cut 200 jobs in US Ads and Android citing AI alignment; Salesforce eliminated about 1,000 in marketing and non-technical functions as it deploys Agentic AI; Meta forecasts 5% of workforce will be cut to double down on AI, and IBM plans to replace roughly 30% of its back-office roles (around 7,800 positions) with AI within the next five years. Or is this the latest excuse for cost cutting in already struggling business areas with the restructuring cart before the AI product horse?
Think this is only tech companies? Amazon expects its workforce to decline over time as AI expands within its organization.
These dire predictions highlight the real potential to rapidly wipe out many jobs. How will society deal with a future of 15% or 20% unemployment, or labor force participation dropping from the current low 60% to the mid-50s? Will we become a permanent welfare state?
Past technology waves such as the early 1980s advent of the personal computer and the rise of the Internet took years to play out, however there were new job roles and indeed entire new industries created. The fear is that AI is advancing so fast and the breadth of the roles and industries it potentially impacts could result in a more dire outcome.
However, this is not certain, and other commentators point out the history of technology adoption. An article on LinkedIn, The F-Chart, tracks the proportion of adults in the UK using the internet, from its infancy in the early 1990s to the ubiquitous utility it is today, and the proportion of people working age 16+ over the same period.
While one metric changed dramatically, the one many predicted to be impacted, the percentage of people working, did not. The widespread adoption of the internet did not change the propensity to work. The key will be how rapidly new roles and industries are created to support the new AI augmented world.
To get-ahead of potential issues and the worst-case scenarios of an autonomous AI world, Responsible AI frameworks are being developed. Responsible AI considers the broader social impact of AI and the measures required to align technologies with stakeholder values, so AI can be deployed in a safe, trustworthy, and ethical way.
Another perspective is Augmented AI, or Human-in-the-Loop AI, (HITL) outlined by Dmitry Kon of Access Development Solutions. This offers a practical approach where humans direct, and monitor, with guardrails and policies, AI implementation to ensure sensible adoption, maximizing ROI, while minimizing risks. While AI is great at automating tasks and eliminating repetition, humans remain a differentiator and having front-line people interacting with customers makes a difference. Overly aggressive automation and self-service risks pushing entire workloads onto customers – risking loss of business,
Additionally, AI can fabricate convincing, but false results, hallucination, that without analysis, can easily become false facts. Some argue that generative AI has polluted the world with bad data, and data sets before ChatGPT can be considered safe and clean, anything after, not so much. The concern is that AI models are being trained with synthetic data created by AI models and subsequent generations of AI models may become less and less reliable
If companies do not have good data or invest in data, the foundations of AI, and therefore its outputs, will be flawed, especially in nuanced areas where judgment and experience are critical. In addition, LLMs lack mechanisms for incorporating real-world feedback, an area ripe for innovation. Additional costs in the form of cybersecurity are also not typically factored in.
The reality is that many companies remain stuck in the pilot phase of AI – isolated use cases, unable to scale. And for the most part, AI remains front-end biased – the technology is more prevalent across customer service, sales & distribution.
The progress of technology is never linear. Reports that Salesforce is blocking AI startups from accessing Slack data and SAP is blocking enterprises from extracting customer data from SAP applications raises antitrust questions. Or is it smart business? Data portability and interoperability are critical for AI’s underlying models and there will be conflicts over data ownership. This battle line complicates the potential of AI.
In addition, the delivery of software will fundamentally change from user licenses for rigid, isolated back-end applications, to dynamic interconnected systems where the customer pays for outcomes.
What is missing is a synchronized approach with strong governance and leadership. It is critical not to start with technology, but to clearly define what you want technology to do. This is where the emerging role of a Chief AI Officer (CAIO) can be critical, or why senior management, with a holistic perspective of the business, must espouse a vision. A CAIO can align stakeholders, break silos and drive enterprise-wide transformation with a holistic strategy.
What does all of this mean for the distribution industry? First, keep humans in the loop, be clear about data ownership and usage, consider the role of a Chief AI Officer for starters, and most importantly, staff must be learning and upskilling continuously in AI-related tools.
As always, we appreciate your feedback and comments.
Leave a Reply