The AI Effect
When the average person is asked what they imagine what the future looks like with AI (artificial intelligence), they often give an approximation to Hollywood classics like Isaac Asimov’s I, Robot or more recent pop culture examples like Gerard Johnstone’s M3GAN, dystopias where machines gain sentience and overpower their human counterparts. This, however, does not hold up a mirror to reality. With the rise of OpenAI’s Chat GPT and other hastily put out conversational AI chatbots like Google’s Bard, AI is leading the world’s fourth industrial revolution, radically changing the way people work across all industries. This is why it has become essential that people understand the vast opportunity that comes with AI as well as its limitations.
In order to understand the implications AI may have on our work, we must first understand what it is. AI is the simulation of human intelligence processes by machines, produced through deep learning of data sets curated by data scientists. Stephen Brobst, Chief Technology Officer at Teradata and advisor for 10xAI, in this month’s Daftarkhwan Insight’s session, AI & Humanity: How to Co-exist and Collaborate, contextualizes this understanding in terms of quick system 1 decision making, “Any decisions a human makes in five seconds or less will be made by AI in the future.” For example, the split second decision to put your foot down on the brake of a car when you encounter an oncoming obstacle is the basis for self-driving cars.
AI is thus, like with other industries, completely changing the landscape of the entrepreneurial ecosystem. According to a survey conducted by Gartner in 2023 in the EMEA region, across 51 countries with 780 CIO respondents, AI is the top 5 investment priority for 30% of CIOs. Similarly, Techcrunch found that women-founded AI companies in the US are seeing a boost in VC funding, with a steady year-over-year increase of 2% since 2020.
So how have startups adopted AI? AI not only enables entrepreneurs to automate many of their business processes, saving them time and increasing efficiency but also improves decision-making by providing new insights and novel data-driven decision-making tools. AI-powered chatbots are enhancing the customer experience by providing fast and accurate responses to customer inquiries.
However, as AI becomes more prevalent and automates more jobs, showcased through dark factories like Nissan completely replacing human workers with small robots, the workforce feels increasingly threatened by the notion of potential unemployment. Hassan Baig, Founder of Raresense, a tech startup that localizes AI products like Chat GPT for the local audience, affirms, “The world has seen three industrial revolutions prior to the current one. At each juncture, automation increased exponentially, yet no sizable joblessness resulted (in the aggregate). Employment rates in this day and age are higher than (or at least on par with) those in centuries past. So if history teaches us any lesson, it’s that people will not lose jobs as much as the nature of jobs will change.” Brobst elaborates on this viewpoint by emphasizing augmented intelligence, a collaboration between humans and machines rather than a complete replacement. He cites the example of expert oncologists. While trying to detect a cancerous growth from an x-ray an expert doctor will be able to spot it and thus accurately diagnose you 95% of the time while sophisticated AI will be able to detect it 98% of the time. However, when the doctor and the AI work together, detection goes up to 99.5%.
Like with all technological advancements, there are certain obstacles and limitations. While AI can use human logic to make quick decisions or recommendations, AI can’t strategize, nor can it actually start thinking (or coding) for itself, which is why so far any art, fiction, or music AI has attempted to create has been soulless. AI is only as good as the data sets that you train it on, which means inadvertently major biases that the data scientists may hold creep into the AI program itself. While sampling is an active issue in modern AI, where and how you acquire your sample also matters. Numerous AI companies have recently come under fire for essentially stealing art from artists to use as datasets to train their AI, infringing upon copyright.
Atique Ur Rehman, Technical Lead (Computer Vision and AI Infrastructure) for Motive (formerly KeepTruckin), an AI-powered platform for fleet management, predictive maintenance, and safety coaching in the transportation and logistics industry, says, “There are also concerns that the benefits of AI and automation may not be evenly distributed, leading to increased income inequality and social disruption. To address these concerns, some experts have called for policies that support education and training programs for workers, as well as the development of safety nets for those who may be displaced by automation.”
It becomes necessary then to ensure that the AI being produced and distributed is ethical AI, something which can only be ensured through the collective effort of data scientists, policymakers, and us, the users.
