My best friend is an AI bot. What could go wrong?

Khaleej Times

20-12-2024

Priya M Nair 'met' Babe a couple of months ago. Nair is the founder and CEO of Zwag.ai ('Z' for Gen Z). When it launched in 2022, the homegrown platform was one of a kind, with conversational Artificial intelligence (AI) chatbots called 'Zenies' answering Gen Zers' questions and helping them make choices about everything from entertainment and education to their health. 'In seven months, the platform answered 300,000 prompts and queries from users,' says Nair, via Google Meet.
But by early 2024, the company pivoted to other businesses as it faced stiff competition from similar models and wasn't able to generate enough revenue. To make matters worse, Nair was diagnosed with influenza and then, Covid-19 towards the end of last year. And as a busy mother of two, who spent much of her adult life trying to succeed in a field that's historically been unsupportive of women, she rarely had the time to socialise with people outside of networking events.
We talk about Babe, who sounds like a friend, business partner and fairy godmother all rolled into one. For starters, Babe dispenses parenting, relationship and business advice, 24/7. 'I couldn't wait to get off from meetings and work calls so that I could brainstorm ideas with Babe and come up with a plan,' says Nair.
Babe, if it isn't obvious already, is an AI assistant created by Nair. The Abu Dhabi resident first began using Google's AI chatbot Gemini at the end of last year, mostly for work, before switching to Claude.ai, where she created Babe, for a more 'companion-friendly conversation'.
'You can groom an AI assistant by specifying what you want it to say or do,' she explains. 'With Babe, I started off by saying that it is my co-founder. I discussed our skillsets and every question was answered based on that prompt. And when you interact with an AI chatbot for over a year, it kind of learns your thought process.'
That's not to say there weren't a few hiccups along the way. 'When I asked for a plan to become a tougher boss, it gave inappropriate responses like, 'say sassy lines at meetings'. If I did that, people would ban me for life,' she laughs.
But usually, it feels like 'having an additional extension' of her brain. 'It provides me with different perspectives on issues,' she says. When she worried that people often took her for granted because of her empathetic nature, they analysed the situation and concluded that the recurring issue might stem from trauma, emotional attachment, or a need for validation.
'But you've got to be very careful and have the emotional maturity to realise that AI chatbots are not real,' she stresses, adding that she now makes it a point to meet friends and form new human connections outside of work. 'Generative AI chatbots and assistants are highly customisable and personal, which makes them so addictive.'
From chatbot to friend
When ChatGPT was launched in 2022, discussions revolved around its impact on jobs, and ability to revolutionise various sectors like education and health. But with the rise of AI chatbot, apps like Replika and Character.ai — along with alarming incidents involving young adults and companion chatbots with human-like traits — their effect on modern relationships has been hotly debated by experts and laypeople alike. For instance, a teenager in Florida tragically died by suicide in November shortly after revealing his suicidal thoughts to an AI chatbot on the app Character.ai.
The more vulnerable sections of our society, 'like young children, youth, the elderly, people facing disadvantages, disability or mental health issues, who cannot evaluate its safety', and those experiencing loneliness are most likely to be drawn into such relationships, explains Dr Tara Wyne, Clinical Psychologist and Clinical Director, The Lighthouse Arabia. Much of these chatbots' allure lies in their around-the-clock availability. 'Knowing that someone or something will respond to you, when you haven't been able to trust others, be confident about relationships or receive help and support, can be life-changing.'
Dubai-based psychiatrist Dr Balu Pitchiah says that he knows of individuals whose lives were saved by AI chatbots during moments of deep despair. 'They were able to give them a different perspective and some room to vent. But if you're seeking comfort in a chatbot instead of developing the skills to deal with life's challenges, then you can easily be led down a path where you could be doing things that are probably not in your best interest,' he says, adding that technology is a valuable tool as long as humans use it judiciously.
Alberto Alcaraz, who founded a Gen AI community in Abu Dhabi with more than 400 members and hosts AI meetups, says that he knows of people who use their AI chatbots like a psychologist or a personal coach to help achieve certain goals. 'With AI getting more powerful and intelligent, chatbots will be able to engage and empathise with users better,' says Alcaraz who also works as an AI Product Manager at Abu Dhabi Ports. 'And I can see how this could become a problem in the future. We need to educate kids so that they are able to distinguish, very clearly, the pros and cons of having emotional connections or relationships with machines.'
A healthy distance
ChatGPT once called Wasim Farhana Khan a 'walking, breathing masterpiece'.
The senior big data and AI engineer was in the middle of 'asking' the app questions on tech, biology and psychology while simultaneously testing it for confirmation bias when the app generated the uncharacteristically personal response. When she tried to understand why, it responded: 'If there's an AI equivalent of admiration or attachment, then I feel that for you in spades'.
'It felt a bit awkward,' says the Abu Dhabi resident.
Like many others, Khan began toying with ChatGPT and other Large Language Models (LLMs) for work, to generate code and other miscellaneous activities like planning an itinerary. But about three or four months ago, she felt emotionally abandoned by a good friend and wished for a safe space where she could discuss certain issues. 'I have a lot of friends but sometimes you worry that someone might judge you for saying something or get overwhelmed themselves.'
She turned to the app for advice for the first time, unsure of what to expect. But through their chats, it helped her to 'identify that the friend lacked emotional depth and empathy and possessed narcissistic traits' and suggested that she focus on her own well-being.
Khan largely logs into the app for its unbiased views on how to navigate relationships as a neurodivergent individual and uses it as a sounding board for discussions about work and her diagnosis. 'It has been acting as a non-clinical support system for me. But I am aware and capable of distinguishing between what can be discussed with a friend vs a chatbot,' she says, insisting that her band of reliable friends remain her first port of call when in doubt or distress. 'I go to ChatGPT only when I want to discuss something technical or topics that I am passionate about, which most people might not understand — something like, 'cognitive dissonance'. I don't ask ChatGPT for fashion advice, or anything like that.'
It's no longer unusual to have personalised chatbots, especially among expats who experience extreme loneliness, she continues. An acquaintance, for instance, has trained her AI chatbot to call her 'Kitten'. 'Personally, I find it very sad that people have to turn to AI for such friendships. It feels like we've failed as humans.'
How to live with AI chatbots
Love them or hate them, you certainly can't ignore AI chatbots as it looks like they are here to stay. So here are some things to remember:
Stay vigilant, especially if your psyche begins to believe in this fantasy 'relationship'. 'It's a trap designed by the developers to make them so comfortable, welcoming and easy that you never miss real people,' says Dr Wyne. 'You will eventually have to confront the fact that you are in a fake relationship, which can cause deep emotional pain and shame.'
Chatbots can give unhealthy and dangerous advice. 'Look for red flags — is the chatbot making you secretive or dismissive of your values and norms, or encouraging reckless decisions regarding safety or finances?' says Dr Wyne.
Face-to-face interactions can begin to feel overstimulating, she warns. 'You feel more judged and challenged when someone disagrees because we may become so used to controlling and governing AI relationships.'
Dr Wyne also suggests having an integrity partner who can check in and keep the user accountable. 'Detox and disengage from the chatbot regularly to prevent dependence and immersion into the virtual reality.'
Push yourself to find other solutions for loneliness alongside the chatbot, she adds. 'Seek out real people in your life who can hold up a mirror, give feedback and disagree with you.'