Global executives fear the increased utilization of AI as the industry is still uncertain about the intelligent technology’s full capabilities and potential
With the accelerated pace of digital marketing, enterprises have already begun to make decisions about adopting AI-driven solutions. A recent US Policy report ‘National Security Commission on Artificial Intelligence’ states that Americans have not yet understood the exact consequences of AI adoption on their national security, welfare and economy. The 756 pages report underlines the fact that while AI should benefit the country, it should also defend it against AI’s destructive capabilities.
True, there is undiscovered knowledge that might reveal overwhelming possibilities. Experts also reflect on Open AI’s CLIP and Facebook’s new AI-model SEER. Darkly fascinating, both involve neural network software, which resembles a human brain. Research suggests that large amounts of unlabelled data that is fed into these neural networks would result in extraordinary capabilities.
Top-five accuracy in computer vision means that at least one of the highest classifications matches the image label. According to Stanford’s AI Index, training an AI system to reach 93 percent top-five accuracy took a little over six minutes in 2018. Last year, it only took 47 seconds.
Meanwhile, the system capabilities of generating images are also improving at a rapid speed. A benchmark citation in the AI index reveals that there has been an improvement of 37 percent in the quality of images generated by algorithms in just two years. Confirming the fact are the recent Tom Cruise viral Deepfakes.
With the rapid developments of AI from cutting edge technology to mainstream comes reduced costs. In 2017, it cost a company more than $1,000 in cloud-computing resources to reach that accuracy level. Last year, it only took $7.43.
When technology gets cheaper, it leads to an increase in utilization. According to research presented by KPMG, several industry leaders across sectors claim that AI is being adopted in their business strategies, either moderately or in fully functional capacity. Several brands believe that the COVID 19– led rapid digital growth was the reason for the increased acceleration in the adoption of AI.
The same report reiterates an implicit caution issued by the US Policy report – the dangers of this accelerating intelligent technology and that AI-driven solutions and tools are yet to be appropriately assessed. Many executives are also afraid that AI is being enforced too quickly. The world has adopted it without fully understanding the technology itself and how it fits into their business strategy. From cybersecurity, governance to ethics, addressing these challenges and building fair, responsible AI-driven solutions have never been more critical than now.
The research also reveals that although the number of research papers with ethics-related keywords has grown since 2015, the average number of paper titles with the same words remains low. The AI Index suggests a reason for the same. The governance and ethics around AI will continue to be delayed as there is inadequate consensus about how the technology should be measured or benchmarked. Based on the current situation, it is hard to say if and when AI will truly be under control.