AI at work: unexpected trends in jobs, teams & trust.

In the past 48 hours, the impact of AI on employment became starkly visible. On Monday, enterprise giant Oracle revealed an expansion of its 2026 restructuring plan that will eliminate around 21,000 jobs – roughly 13% of its workforce – and attributed part of this reduction to the adoption and deployment of AI technologies ([1]). The cuts come as Oracle pours billions into AI-driven cloud infrastructure, and they add to a surge of tech layoffs seen in early September. Industry data shows that workforce cuts in the tech sector jumped by 199% in September compared to August, with large internet firms accounting for 77% of the 4,400 layoffs recorded so far this month ([2]).

Yet job displacement is only one side of the story. Paradoxically, the rise of AI is also creating demand for new roles and skills. Companies that win major AI-driven contracts are rapidly expanding their hiring of tech and data talent. For example, one IT services firm, Fulcrum Digital, is adding 800 new technology positions across the US, Europe and India after securing a large enterprise AI deal – including specialized roles in AI solution architecture, machine learning operations (MLOps), and even emerging titles like 'prompt engineer' . This reflects a broader trend: while AI can streamline or even replace certain tasks, it is simultaneously generating new work in areas like AI development, maintenance, and oversight. So far, the net effect on overall employment has been relatively modest, rather than the wholesale human obsolescence some feared. However, early signs of workforce shifts are appearing in specific segments. Recent analysis suggests that entry-level hiring is slowing in roles highly exposed to AI automation – one study found a roughly 19% shortfall in new hires aged 22–25 for such jobs ([3]). In short, AI’s impact on jobs is proving uneven: it’s driving strategic cost-cutting in some quarters while opening new technical frontiers in others.

Even as companies restructure in response to AI, many are simultaneously redesigning work itself around AI capabilities. This week brought clear evidence that digital 'co-workers' are moving from pilot projects into core business operations at a global scale. On September 15, enterprise software leader Salesforce announced a new portfolio of job-ready AI agents for functions like sales, customer service, commerce, HR and IT . These agents come with pre-built skills and actions, connect directly into Salesforce’s Customer 360 platform, and can even pursue long-term goals over days or weeks using a new 'long-horizon' AI runtime . In practice, an AI agent might autonomously perform multi-step tasks – such as identifying and re-engaging at-risk sales prospects or handling parts of employee onboarding – while human colleagues supervise and step in as needed. Notably, the global staffing firm Adecco Group has signed on as an early adopter and will deploy Salesforce’s Agentforce AI coworker across 40 countries to assist recruiters with candidate pre-screening, hiring and onboarding at scale ([1]).

A similar embrace of AI-driven teamwork is happening in other regions. In China, tech giant ByteDance (best known as owner of TikTok) just rolled out a major update to its Feishu workplace collaboration app, embedding AI agents deep into everyday workflows . With Feishu’s millions of users, these virtual assistants – capable of automating routine coordination tasks in chats, scheduling, and project management – could rapidly become an ordinary part of office life, effectively normalizing AI 'team members' for a vast number of workers . By building AI directly into a leading collaboration platform, ByteDance is likely pressuring other enterprise software vendors to follow suit.

These developments show how forward-looking organizations are not merely using AI to accelerate existing processes; they are re-architecting roles and teams to integrate AI as a collaborative partner. Some, like South Korea’s Hancom, are even explicitly presenting AI as a new kind of digital workforce. This week Hancom introduced Nomadian, a platform that lets managers assemble virtual teams of multiple AI agents – each with defined job roles – inside a 3D office environment ([2]). While playful in presentation, the concept underscores a serious point: companies are exploring ways to have AI systems collaborate like human teams, complete with role assignments and built-in checks such as required human approvals for final decisions ([3]). In effect, AI is graduating from a set of experimental tools to an active, integrated part of the workforce.

Amid this rapid infusion of AI into the workplace, how are employees reacting? New research suggests that workers are both enthusiastic and anxious. According to a fresh Gallup workforce survey, 52% of U.S. employees now report using AI in their day-to-day jobs – up from 50% just a quarter earlier ([1]). Many workers say these tools bring personal benefits: 66% of AI users report that AI lets them spend more time on high-value tasks, and 58% say they are accomplishing work they couldn’t have done a year ago without AI’s help . Far from replacing human ingenuity, AI is often taking over drudge work and enabling people to focus on higher-level, more creative responsibilities.

Yet alongside this optimism comes significant undercurrents of concern. A recent Pew Research Center study found that 71% of Americans believe AI will result in substantial job cuts across the economy over the next 20 years ([2]). Whether or not those long-term fears prove justified, a more immediate issue is clear: many employees don’t fully trust how AI is being implemented in their own workplaces. In one global survey, 57% of workers admitted they hide their use of AI or present AI-generated output as their own work . This kind of 'shadow AI' behavior often stems from uncertainty – employees fear how managers might react or are unsure what’s permitted. In fact, more than half of workers in that study said they have used AI without their employer’s knowledge or formal guidance, and a similar proportion reported mistakes in their work due to unchecked AI outputs .

There are also signs that AI-driven changes can affect morale and culture if not managed well. Among employees at organizations introducing AI, about one-quarter say their workplace culture has worsened – roughly the same share that say it has improved . This split suggests that when AI adoption is handled haphazardly or without employee buy-in, it can breed frustration and fear, even as others find ways to thrive with the new tools.

All of these trends point to one conclusion: the role of leadership is now central in determining whether AI’s disruption becomes an opportunity or a threat for the workforce. Many executives know the stakes. In a recent poll of Fortune 500 chief human resources officers, 99% affirmed that AI is important to their company’s strategy – but only about half were confident that their managers have the skills to guide employees in using AI day-to-day . That caution is mirrored on the front lines: only 26% of AI-using employees feel their top leadership has articulated a clear plan for integrating AI into their work . This disconnect leaves a majority of people in the dark, potentially breeding the very fears and missteps – from job insecurity to secretive AI use – that we’re now seeing.

Forward-thinking organizations treat AI adoption not as just a technology project, but as a comprehensive change-management challenge. They are investing in training, communication, and culture – not just software – to help their teams thrive alongside AI. Many have launched internal AI academies or cross-functional 'fusion teams' to boost skills, aiming to upskill a broad swath of employees rather than a few specialists. Early evidence suggests this people-first approach pays off: companies unlocking the most value from AI tend to be the ones empowering their workforce to implement and innovate with it. Experts note that the algorithms and tools themselves account for only a fraction of AI’s potential value; the bulk comes from reimagining business processes and roles to fully leverage these technologies. In practice, the more an organization enables its employees to learn and experiment with AI in their roles, the greater – and more sustainable – the performance gains.

Perhaps most importantly, effective leaders model the mindset they want to see. When managers visibly embrace AI in their own day-to-day work and encourage their teams to do the same, employees’ perceived benefits from AI can rise dramatically – one study showed a 17-point jump in reported value when leadership actively used AI tools themselves . Likewise, employee trust in AI systems increases by around 30 percentage points under leaders who are transparent about how AI is used and what its limitations are . Leaders who communicate a clear vision for AI, set guardrails for responsible use (in consultation with HR and even worker representatives), and celebrate examples of AI augmenting – not replacing – human potential will cultivate a culture of trust and learning. Those organizations will be far better positioned to navigate the upheavals of AI without losing their people along the way.