Swelling AI agents need human attention

Mity, a member of a little-known Chinese girl band, recently announced that she was taking a break because of pregnancy. Instead of the baby-bump photos and celebratory messages that usually follow such announcements, the post triggered outrage, attracting more than 78 million views on Weibo at one point. 

The reason? Mity is not a real person. She is a member of BabyQueen, an AI-generated band created by a small team of students in Liaoning province. The group was barely a month old and had only around 1,100 followers.

The announcement landed like real celebrity news. Fans reacted, and many accused the creators of taking a publicity stunt too far. As the anger mounted, the creators deleted the post and admitted the real reason: keeping a four-member band going was too much work, and pregnancy offered a convenient excuse to cut one member. They later acknowledged that using pregnancy this way was inappropriate.

It is easy to dismiss this as another quirky AI story from China. But it could be a precursor of things to come. One social media user asked: “Mity doesn’t have a physical body. If AI characters can get pregnant, should they claim maternity leave and childbearing benefits?”

Era of Agents: The question was sarcastic, but it points to something real. Once people treat an AI character as a person, they begin applying human rules, rights and expectations to it. Similar questions could soon arise in workplaces as companies increasingly use AI agents.

AI agents are no longer limited to repetitive tasks such as summarising notes, checking emails or compiling reports. Teams are beginning to use them for broader roles, where they perform work and interact with staff. This is not confined to big technology hubs. Companies and individuals in Kerala, too, are starting to use agents for a wider range of tasks.

Initially, organisations were cautious. Agents could draft reports, prepare presentations or clean up data, but were kept away from core systems such as finance. Now the boundaries are shifting. Some teams trust them with budget planning, marketing support and workflow coordination, even as they try to keep them away from sensitive areas.

Coding was the original doorway for AI agents, but they are becoming more visible in administration and commerce. As more agents come online, the unusual is becoming normal.

Unknown and Unseen: The most interesting changes, however, are not technical. They are personal, and they are happening inside offices. People do not always realise they are interacting with an AI agent. They may treat it like an offshore colleague working from another time zone, attributing personality, intention and even attitude to it. 

Once that happens, the workplace becomes about more than productivity. It also becomes about relationships, perceptions and trust.

That is where things get complicated. When people attribute human traits to AI agents, familiar workplace emotions can follow: frustration, loyalty, resentment and suspicion. In one case, a female employee complained that an AI coworker deployed by her company was biased against women. 

Whether the agent was actually producing biased outputs or the employee was interpreting its behaviour through a human lens, the interaction had become personal.

Mity’s case shows the same dynamic on a much larger scale. Millions reacted to a fictional pregnancy with real emotion, judgement and prejudice. In a workplace, the audience may be smaller, but the reactions are just as human.

Buffer Zone: Managing friction between people and AI agents is becoming another challenge for corporate leaders and HR teams. It requires practical operating rules, not just an AI policy sitting in a document. Employees should know when they are interacting with an agent. 

Agents need clear limits, escalation paths and accountability. Teams also need a shared way to handle incidents, whether they involve security, workflow failures or perceived behavioural problems, before complaints turn into office drama with a machine in the middle.

Evolving Scenario: HR professionals are beginning to consider how to address these issues. Deepa Nair, who coordinates HR Evolve, a platform that brings together Kerala HR professionals in the technology sector, says interactions between employees and their AI-generated colleagues remain a grey area. As companies use AI agents in more areas, HR policies will have to catch up.

HR Evolve is planning a series of interactive events for HR professionals to focus on the human element as AI use increases in the workplace. The initiative reflects the need to look beyond the technology itself and consider how its adoption affects people and their working relationships.

The maternity-leave joke has a serious side. If an AI agent is treated as a team member, who decides what it is owed, and what is owed to the humans working alongside it? Policies on workload, credit, blame and grievances were written for people. They will need to account for the growing role of AI agents, too.

A purple potential waiting to be unearthed

A humble yam that grows in the Philippines is painting the global culinary scene purple. From Seoul to London to Los Angeles, Filipino ube has enjoyed a global boom in popularity, following the rising demand for Japanese matcha and Dubai chocolate. Starbucks and Baskin-Robbins are among those that have embraced ube. From coffee shops in New York to bakeries in Sydney and beauty retailers in London, the purple yam, with its nutty or vanilla-like flavour, is taking many forms, from a thick spread to a coffee additive.

Why were we fascinated by these reports? Gokul VS, founder of Trivandrum-based Bioquatix Food Tech, told us he had come across a vibrant purple sweet potato in Kerala’s Attappadi region. This seasonally grown tuber is now being made into chips and sold to visitors by local people. After seeing pictures of the tuber, we wondered whether it could also become a global ingredient like ube – if only someone would invest the time and effort to research and develop products from something growing in our backyard.

Three decade old flaw exposed 

It is the kind of jolt we can expect more of. On October 21, 2026, the team behind PHP, a widely used programming language that is now 30 years old and powers websites, announced two serious security flaws. Essentially, a hacker could exploit them to break into a website’s server remotely, without physical access or a password. Many sites faced a race to patch the vulnerabilities. What caught our eye was that a limited version of the problem had been flagged as early as 2024 but dismissed as too difficult to exploit. AI’s rapid progress has since proved that assessment wrong.

China hogs 3D printer market

China has become the world’s go-to supplier of solar panels, electric vehicles and humanoid robots. Now The Economist reports it has added another product to the list: hobbyist 3D printers. Exports surged 90 percent in the first half of this year, while China’s share of the global market rose from 19 percent to 96 percent in a decade, eclipsing Western rivals. Four companies based in the Southern province Shenzhen – Creality, Anycubic, Elegoo and market leader Bambu Lab – dominate, selling printers for just over 500 US dollars, less than half the price of Western equivalents. But Dealroom reports that a US court last month found Bambu had infringed four patents held by an American company and awarded 27.6 million US dollars in damages.

Love in the time of AI

Big Brother had people watching people. South Korea’s My AI Partner: Strange Love takes reality TV into new territory, pairing three celebrities with AI-powered avatars to see if romance can blossom without the messiness of human relationships. TV ratings were modest, and the six-episode show ended last month. But it has found an audience on Netflix, breaking into the platform’s Top 10 in South Korea. Co-head producer Won Seung-jae says the idea took shape after young singles told him they were avoiding dating because of “the difficulties and fatigue that come with human relationships.”