Has Generative AI peaked? asks Oguz A. Acar in HBR. He suggests that when generative AI models recursively train on their own outputs, they enter their own echo chamber, leading to Model Autophagy Disorder (MAD). A human touch is critical, and if we do not find a way to infuse this, Oguz argues that we might have already seen the pinnacle of generative AI — it might now be on a decline. Indeed, Gartner placed Generative AI on the Peak of Inflated Expectations on the 2023 Hype Cycle for emerging technologies. And we all know what comes after the peak: the trough of disillusionment. That said, Gartner also expects the next 2-5 years will bring transformative change to organizations that deploy this technology.
The executive order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence signed by President Biden on October 30, 2023 aims to advance and govern the development and use of AI in accordance with eight guiding principles and priorities. The directive will call for heightened transparency from AI firms concerning the operations of their models, and it will establish new guidelines, especially focusing on the labeling of AI-generated content. Furthermore, the directive will put into action a range of fundamental measures related to evaluation, monitoring, and risk management, in accordance with the guidelines set forth in the Blueprint for an AI Bill of Rights and the AI Risk Management Framework.
According to Fei-Fei Li, one of our most prominent computer science researchers, AI is at an inflection point. Li argues that Gen AI has increased awareness of AI implementations for everyone using tools such as Chat GPT. And businesses are increasingly using these tools to help them better serve their customers and drive their profitability.
Furthermore, the talent war is in full swing. Open AI is trying to attract some of Google’s best researchers with compensation packages between $5 million and $10 million. This is a sign that Open AI is gearing up for yet more ambitious products that will further disrupt how we work. After their first developer conference, OpenAI GPTs disrupted several start-ups and aimed to disrupt the open-source community with their functionality and pricing.
So, which is it? Are we in decline, as Professor Acar argues? If not a decline, at least a slowing in the adoption driven by organizations taking their time to decipher and implement the priorities of the Executive Order? Or are we at an inflection point, as Dr. Li suggests? The data on hiring and activity in this space certainly indicates a period of significant growth. Can both be true?
Honestly, it does not matter. Arguments suggesting the end or the beginning are good to follow — at least to understand the perceived limitations and endless possibilities of Gen AI. And it is natural that new technologies will go through spurts of growth and pause. Instead of getting caught up in these cycles of hype and disappointment, we should focus on the use cases that drive the most outstanding value now. The existential threat that AI poses to humans is an exciting water-cooler discussion, but we have more immediate problems to tackle and solve.
Organizations are currently deploying generative AI models to
· Query and generate insights from a large corporate data corpus (product marketing, technical documentation).
· Use AI-generated marketing copy and summaries and combine this with experimentation to drive open and click rates.
· Increase L1/L2 customer representative productivity with a Q&A assistant that answers product-related questions and addresses customer concerns.
· Accelerate lead acquisition and conversion with an AI assistant that identifies opportunities and automates content workflows (sharing relevant content and building relationships with leads).
· Automate code restructuring and optimization while simultaneously generating comprehensive documentation.
· And many more.
While many of these examples augment existing processes with Gen AI, we are looking at the near-term horizon to transform functions.
This technology is impressive in what it can do. But most compelling for me is its ability to bridge the skill gap. Studies have shown that with Gen AI, low-skilled workers can achieve performance close to that of high-skilled workers in many tasks. And in a world where technology ensures that we can all do what we do better, I am sure we will direct this energy to making the world a better place.