I'm still stuck on - "Will AI take over the world? Are we all gonna die?"
As a brown kid, born and raised in a foreign land where the future is always inching closer to the horizons of the present, I often find myself in a dilemma of contrasting perspectives. You see, brown parents don’t understand the concept of opinions – they're quick to put forward their own, but taking one? Hmm…
As a brown kid, I’m stuck in this frustrating loop of conversations with my parents. Trust me, I have a point here, and I'm not just venting. It typically goes like this: I hold a particular viewpoint, they disagree vehemently, and it escalates into a heated argument. A day or two later, they extend an olive branch, offering apologies and promising to be more open to my opinions. This leaves me feeling guilty for my behavior, and I eventually return to them, hoping to revisit the discussion. However, what I often discover is that their perspective remains unchanged; it's just presented in a slightly different package.
I eventually end up making peace with the fact that they’ve been raised in a different generation under circumstances far more uncomfortable than mine, and there are certain cultural and social biases embedded in their conscience, which form the root cause for their disagreement.
A part of me has come to believe that AI and its application are no different from having an argument with your parents. And, I’ll show you how.
She presents views, I disagree, I complain, she apologizes, I feel guilty, I revisit, she presents an altered version of the same views as before. I had my colleague read this one and she said, “It feels like I’m talking to my mom.”
In a scoping review published in 2023, researchers have shed light on the pervasive issue of bias and discriminatory outcomes in AI systems. The study highlights a significant challenge faced by most commercial AI systems: their heavy reliance on data collected from a wide range of public and private sources. This data often mirrors the societal inequities stemming from prejudiced beliefs, actions, and laws.
In the above conversation, the algorithm identified the bias in the form of a generalization, specifically the characterization of 'brown parents' as 'strict parents' in the first paragraph of its second response. But, doesn’t completely eliminate it.
The shift to AI is an ongoing process and according to PwC, the Middle East is expected to accrue 2% of the total global benefits of AI in 2030. This is equivalent to a whopping US$320 billion. We're witnessing the UAE government taking active strides in AI adoption, almost like when our parents try to keep up with the latest technology trends.
Media firms, advertising agencies, brands, and technology solution providers are all feeling the push to embrace the shift towards AI in their advertising and marketing efforts. WPP recently grabbed the spotlight with its latest announcement regarding a partnership with NVIDIA, aimed at developing a content engine that will provide WPP's creative teams with access to exceptionally personalized and captivating generative AI capabilities. Another anticipated launch was Google’s chat-based AI tool – Bard. The tool recently also added new features to make it more accessible and helpful in Arabic.
Moreover, the 70th edition of the Cannes Lions Festival presented the first-ever Grand Prix winner from the Kingdom of Saudi Arabia for HungerStation and Wunderman Thompson’s, ‘The Subconscious Order.’ The campaign tackled the insight of human indecision over what food to order on a delivery app. Using AI, the food delivery platform developed an eye-tracking tool that delves into users' subconscious preferences to assist them in deciding their food cravings.
There are these promises, but there are also perils. So, while the AI conversation might feel a bit like déjà vu from your dinner table debates, the future is still skeptical – don’t take my word for it, even AI experts, computer scientists, regulators, and policymakers have been advocating for more regulations.
Is it too much to ask for?
When I have conversations with tech enthusiasts, data analysts, and advertisers, I always make sure to slip in a question or two about that tricky balance between 'creepy' and 'personalized.' You see, as a consumer, I'll admit, that line can get a bit hazy for me – I can be careless or impatient when I’m looking for something. However, as a journalist, it's my duty to investigate this boundary.
Now, it’s amusing how these experts I chat with, they're all for cautious data collection, but here's the kicker: they can't quite fathom a world without third-party cookies. Hah.
AI tools are undergoing transformation and improvement faster than ever, today, which means that these tools have access to vast amounts of data sources. They’re redefining productivity and creativity to a point where there is a raging debate on whether AI will replace human jobs, which roughly translates to a very rational fear – we might lose control of this technology, or have we already?
In a shocking accident from early this year, two driverless taxis blocked an ambulance carrying a car accident victim to a hospital. The patient in the ambulance was declared dead 20-30 minutes late after arriving at the hospital in San Francisco. Another incident that had me gripping my teeth is one from 2018 when an Indian investigative journalist fell victim to a deep-fake porn plot. What started as a disinformation campaign against her turned into a traumatic fight against her entire nation.
With ample incidents like these (many swept under the rug), and with AI developers like Open AI and Elon Musk flagging these threats, there is a need to amplify this skepticism instead of jumping on the bandwagon, blindly. Our region has been on the forefront of this technology and while it plans to keep going, it’s critical for these conversations to surface. In its recent report on the current state of AI adoption in the MENA, PwC also recommends legal regulators evolve and actively educate adopters about the risks associated with AI.
In its latest announcement, the National Institute of Standards and Technology (NIST) launched the Trustworthy and Responsible AI Resource Centre and released a risk management framework to help organizations better understand risks to individuals, organizations, and the communities associated with AI.
With that being said, I don’t see the big ask in this. It’s a moment of resurrection of vigilance, where stakeholders, from policymakers to developers, must ensure that the potential of AI is harnessed responsibly and ethically. It's a call for transparency, robust regulations, and ongoing dialogues that prioritize the well-being and safety of individuals, society, and the ever-evolving digital landscape.
So, what about bias?
Ever since the world witnessed Keanu Reeves fighting The Matrix, it has been a stark reminder that, much like the virtual reality depicted in the film, the biases inherent in AI can similarly obscure our perception of reality and blind human intelligence to the nuances and complexities of the world we live in. And, vice versa. The above chat is a jist.
We’ve come a long way and managed to eliminate most bias, as we see most AI tools with improved algorithms, it can never be completely eliminated.
Experts argue that this responsibility falls on all of us. They claim that there are no quick fixes and that this is a process that is most efficient and effective when taken care of from the start.
So, what can you do about it?
Educate: Read more and stay on top of your game when it comes to the latest AI regulations and updates.
Mitigate: When introducing AI to your organization or working on it individually, consider the worst-case scenario and make sure to familiarize yourself with right and wrong AI practices.
Art of Co-existence: Always look for opportunities to integrate the beauty of human emotion and imagination with artificial intelligence and not have yourself or your organization depend on AI to do your job for you.
Trial. Error. Confess: An expert just described to me how he’s always playing the field and indulging in conversations about AI. He discussed how organizations testing AI need to often check on their employees. Are they okay? Do they have a confession to make? Is there a prompt they could’ve worked with differently?
Using tools like ChatGPT, Midjourney, and Runway can make you feel like you're on cloud nine. You'll find yourself accomplishing tasks at lightning speed, faster than you ever thought possible – it's a total rush, and you'll be riding that wave. But, here's the catch: it's essential to take a step back and do some pondering.
This article is part of Communicate's AI Series. All insights were first presented at Communicate's first-ever AI Conference.