THE AI DILEMMA; WHERE DO WE GO FROM HERE?
As an innovation enthusiast who's always chasing the next wave of improvements to help our clients and business, I've been pondering this week about the thrilling world of generative AI. While I'm enthralled by the potential transformative impact this technology could have on our lives and businesses, I can't ignore the growing pangs that this nascent field is currently experiencing. As much as I want to plug into a machine and find all the answers tailored to my unique needs, it seems a little whimsical at this stage.
Case in point, an incident that recently rocked the legal world. A New York attorney found himself in court after his firm utilized the AI tool ChatGPT for legal research. The attorney’s court filing included references to non-existent legal cases. He confessed to the court that he was "unaware that its content could be false." It's a stark reminder that while AI is undeniably powerful, it's not infallible.
In an insightful conversation I came across, former IBM CEO, Sam Palmisano, labeled AI as a transformational technology, albeit at an early stage. He believes that the regulation of natural language models still needs to take shape. Constructing such models necessitates colossal infrastructural investments including energy, chip power, and processing of massive datasets. Despite the challenges, there's a silver lining. AI could usher in remarkable efficiencies across all sectors.
However, a critical issue that needs addressing is the AI's self-checking or cognitive ability. The thought-provoking question that lingers is whether the AI models can embed self-checking capabilities. At present, this feature is non-existent, posing a significant dilemma as the governments deliberate on regulatory measures. As Palmisano pointed out, it’s hard to regulate something when you’re unsure of its nature but apprehensive about its potential.
The regulatory conundrum aside, one must not ignore the transformational potential of AI. Can we balance regulation without stymieing the innovation and the massive societal impact this technology can usher in? For instance, AI can help bridge the gaps in the quality and consistency of decisions in sectors like healthcare and education. Sam's take on this was compelling: propose guidelines, not restrictive rules, and let innovation blossom.
Pioneers in the field seem to have already embarked on this adventure, leveraging AI to create opportunities. Many US investment banks, for instance, are heavily investing in AI to develop chatbots for financial advice. This step might seem a bit daring given the prevalent uncertainties. However, these financial institutions are demonstrating their trust in AI's potential.
A standout example is Morgan Stanley's Jeff McMillian's vision for deploying generative AI. He proposes combining generative AI's large language models with a company's intellectual capital and knowledge base. For instance, Morgan Stanley aims to integrate their research department, product experts, and a knowledge base of over 100,000 into an AI-driven ecosystem.
The ultimate goal is to provide accurate information with no AI-generated "hallucinations." All content and answers would undergo validation by an independent entity, ensuring transparency in the responses. The system would not only deliver the answer but also the rationale behind it, directing the user to the source of information.
A unique aspect is the development of a feedback loop where the answer "I don’t know" is acceptable. A specialist from Morgan Stanley would step in, resolve the query, and the AI system learns for future reference.
This approach leverages AI's capabilities within a controlled ecosystem where input and fact-checking are validated and rigorously triangulated. It appears Wall Street has swiftly recognized AI’s potential and has begun to leverage the early-stage opportunities it presents.
In conclusion, generative AI is indeed a game-changer, with its fascinating potential and equally intriguing challenges. As we grapple with understanding and regulating it, pioneers are charting the course, offering glimpses into the transformative future that AI promises.
For further reading see the article links below:
https://www.bbc.com/news/world-us-canada-65735769
https://www.bbc.com/news/technology-65789916
https://www.cnbc.com/2023/05/25/jpmorgan-develops-ai-investment-advisor.html