Should you be Worried about Artificial Intelligence?

Short answer: no. Long answer: It's complicated
Published: 29 May 2023

Having dedicated my career to funding, building, and growing new technologies, I’ve been present for many disruptions. In the post-dot-com bubble pop era, we witnessed the revolution of communications, online payments, marketplaces, advertising, and media by companies like Skype, Amazon, Uber, Google, and Netflix. These companies leveraged new technologies to build new business models or digitise ones that, in some cases, had existed for hundreds of years. Long after Facebook became a verb, we witnessed the eruption of Web3 and the eruption of evangelist techbro social media influencers shilling their special brew of lifestyle grandstanding and NFT minting.

But, as a former venture capitalist, I’ve yet to see a nascent technology so wholly and rapidly evangelised outside of its own circles of investors, builders, and users as Artificial Intelligence. It’s almost as if Steve Jobs rose from the grave to deliver a keynote on the potential of AI. So, out of curiosity for Steve’s insights on the topic, I tasked OpenAI’s ChatGPT tool with bringing him back to life and introduce AI in the style of one of his famous keynote speeches:

Good morning everyone, and welcome to this exciting keynote on artificial intelligence.

Today, I want to talk about a revolutionary technology that has the power to change the world as we know it. It’s a technology that’s been around for decades, but it’s only in recent years that we’ve seen its true potential start to unfold. I’m talking, of course, about artificial intelligence.

AI has the power to automate many of the tasks that we do every day, freeing us up to focus on the things that matter most. It has the power to make our lives easier, more efficient, and more connected. But more importantly, it has the power to solve some of the world’s biggest problems.

Think about it. With AI, we can tackle issues like climate change, poverty, and even the spread of diseases. The possibilities are truly endless, and the potential for impact is massive.

At the same time, we need to be mindful of the potential risks that come with this technology. We need to ensure that AI is developed in a responsible and ethical way, and that it serves the interests of humanity as a whole.

So let’s get to work. Let’s harness the power of artificial intelligence to build a better future for all of us. Thank you.

AI Steve Jobs Speech generated by ChatGPT

As AI Steve highlighted above, AI itself is not new. Prior to the release of tools like DALL-E, we saw AI leveraged for specific use cases across most major industries such as marketing, cybersecurity, and even CGI / animation in media. We’ve been using the technology for decades to classify, analyse, and create data (including text and images) for narrow sets of tasks, which is why it’s referred to as “Artificial Narrow Intelligence”. In contrast, new models allow for many use cases with no additional training or fine-tuning. This evolution from the previous generation of AI to today’s Generative AI models underpinning applications such as ChatGPT, DALL-E and others has been driven by advances in computing power, cloud data storage, and machine learning algorithms.

Unlike Web3, AI has already demonstrated its usefulness and potential beyond theoretical adoption. Also, unlike its predecessor in the timeline of popular new technologies, it doesn’t require mass adoption of a new protocol or regulatory approval. There are two possible applications of AI: the optimization of existing digital processes and the digitization of human tasks.

Generated with DALL-E | Prompt ‘Dubai in the style of Edward Hopper’

Optimization increases the speed and reduces the need for human input into existing algorithms. A straightforward example would be chatbots, which had their moment in the latter half of the 2010s and are making a comeback, armed with better-trained algorithms. Chatbots trained with existing customer care data sets will replace notoriously difficult-to-navigate FAQ pages on websites and costly call centres. The result will be a lower cost to do business and improved satisfaction.

This brings us to the frightening or exciting – depending on who you ask – scenarios where AI leads to the replacement of human roles. Short-term, this could range from copywriting, software engineering, art, animation, business analysis, and journalism. But, again, this isn’t a futuristic dream pontificated by the Silicon Valley elite on their three-hour podcasts; this is happening today. For example, Buzzfeed recently announced that it would start using ChatGPT to write journalistic pieces after technology journalism outlet CNET was found to use the same tool to write personal finance articles.

Readers should consider that despite the countless applications of AI, there is no cause for alarm when it comes to making the human worker redundant. The evolution of technology is an inevitability and we are better served by preparing for it rather than resisting it. Many pundits draw comparisons to the widespread fears of the industrial revolution replacing jobs. In the short term, these fears are unfounded. So long as the models underpinning applications exist in a state of ANI, even the most advanced tools will require human input and oversight. However, these tools will complement and augment human work by replacing menial, repetitive tasks in creative and technical fields. For example, this article was reviewed using Grammarly to check for spelling and grammar mistakes. 

Although we’re progressing beyond ANI, there’s still quite the journey ahead of us and little consensus on when we might reach our destination. Some scientists estimate that we’re decades away from progressing to the next state of Artificial Intelligence: Artificial General Intelligence (AGI). AGI would offer capabilities such as sensory perception, advanced problem-solving, fine motor skills and even social and emotional engagement. There’s quite a distance to travel from writing Shakespearean sonnets about lost socks in the dryer to developing a personality like that of Samantha, the protagonist’s AI companion in the 2013 film Her. It’s impossible to predict how soon we could begin to describe AI as AGI; estimates on timing range from a decade or twenty, if ever.

When it comes to Arabic, language models need to catch up. Today’s models are predominantly trained on content publicly available on the internet: webpages, Reddit, and Wikipedia make up approximately 85% of ChatGPT’s training data set, for example. Considering that approximately 60% of written content online is in English and less than 1% in Arabic, the inputs necessary to achieve the same quality outputs in the latter are nonexistent. It’s no secret that English is the lingua franca of Middle East business. Still, we should ask ourselves whether this will further subdue the use of Arabic in such settings. The impetus to ensure the development of the Arabic language in technology and business settings lies both in the private and public sectors in wealthy Gulf countries.

While there are reasons to celebrate AI’s coming of age, we need to keep our feet on the ground. The limitation on the applications of AI is compounded by questions of ethical standards, reliability, accuracy, and truthfulness raised by academics such as Gary Marcus, a leading AI sceptic. Even Mira Murati, the CTO of OpenAI (creators of the models underpinning DALL-E and ChatGPT) is arguing for regulatory oversight of AI. Questions remain on how to solve topics such as moderating offensive model outputs, intellectual property infringement, policing disinformation, and academic honesty to name a few. 

“How do you get the model to do the thing that you want it to do, and how do you make sure it’s aligned with human intention and ultimately in service of humanity?” 

Mira Murati, CTO, OpenAI

Make no mistake, AI is beyond the point of no return, but that doesn’t mean we can’t harness its power to empower our workforces and transform our lives. Although the excitement surrounding AI’s potential is justified, the challenges of its usage and misuse are much more significant than those of previous generations of technology and should not be taken lightly. We have at our disposal an incredible new tool; however, we must balance our eagerness to watch with mindfulness of the risks and implications and careful regulation. 

Rayan Dawud is a former venture capitalist who has held senior roles at Careem and Outliers Venture Capital in Dubai. He’s currently on a career break in London, where he’s exploring Artificial Intelligence.

Featured image generated using DALL-E with prompt ‘android from the film Ex Machina in a Hopper painting’

Originally published on Esquire ME

related posts