#7: The Future of AI .......and more
Deepak Seth
Actionable and Objective Insights - Data, Analytics and Artificial Intelligence
What happens when 2 Futurists (one real and the other a wannabe) get together?
They try to out-predict each other so much that they end up inventing a time machine just to settle the argument! (Thanks?#chatgpt?for the wacky humor)
But jokes apart (no we did not invent a time machine, atleast not yet) , had a blast talking to?Kevin Benedict?about some serious stuff:?#generativeai. We traverse quite some ground - the past (how it came about), the present (where it is now) and the future (where is it headed next) all juxtaposed in context of our personal journeys using it and figuring it out.
We talk about the value it brings as well as the perils and pitfalls. The ethics and equity aspects, need for regulations and controls, the balance between unbridled enthusiasm and appropriate guardrails. End with a piece-de-resistance - "Human Brain augmented AI"
Check it out. Love to hear your thoughts and feedback.
Youtube:?https://lnkd.in/gW33VAiM
Podcast:?https://t.co/fKlVzyWcqk
One of the questions that I cover: Where can Generative AI be used in context of the large enterprise?
The impact of large language models like ChatGPT and generative AI platforms is vast and has the potential to revolutionize various aspects of enterprise operations. From a productivity standpoint, these tools can serve as productivity enhancers for employees across different roles and departments. By embedding them in software like Microsoft Office, salesforce et al as "copilots" employees can have better access to information and improve their research capabilities.
One significant area where these tools can make a difference is in customer interactions. For example, call centers and customer service representatives can leverage these models to quickly access relevant information and provide accurate and concise responses to customer queries. Similarly, field workers who need to refer to multiple manuals can benefit from these tools by obtaining precise information on equipment repair or maintenance.
Furthermore, large language models can assist with tasks such as generating legal briefs, creating drafts, and providing guidance on various subjects. They can act as a handy tool for learning and problem-solving, just like people use YouTube to acquire knowledge in different domains. For instance, using ChatGPT to seek advice on home maintenance issues, which provided you with valuable information and helped you approach contractors more confidently.
However, when it comes to large enterprises, there are concerns about data privacy and the potential leakage of confidential information. Companies want to ensure that their salespeople or employees don't input sensitive client data into these models, which could then become part of the model's knowledge base. This concern arises because the models are trained on historical data, including public information, and there's a need to establish guardrails to protect proprietary information.
Additionally, there's an interesting future perspective to consider. As these models continue to be used extensively, they may start generating their own content, gradually replacing existing public domain information. This raises questions about biases and the potential amplification of existing biases if the models learn from biased data and reinforce those biases in their outputs.
To address these concerns, some discussions have emerged around using technologies like distributed ledger or blockchain to secure and authenticate confidential information within organizations. This approach aims to allow employees to leverage the benefits of large language models and generative AI while ensuring that confidential data remains within internal boundaries and isn't shared publicly.
We did also talk about the?creative possibilities of?GenerativeAI?and the compensation of training AI?with?Artists' and Writers' original?work. This was a fascinating part of our conversation . Brought back memories of some crystal ball gazing about what I had then called "Digital exhaust royalty payment": Has the time come now for users to be paid royalty for use of their “digital exhaust” by corporations? Will a digital exhaust royalty framework be win-win for both: the corporations whose business model is based on the monetizing of this exhaust; and the individuals who are increasingly wary of this exhaust being put to wrong use without their knowledge or permission?
On the call Kevin has me put my futurist hat on:
Where do you see AI going in the next five to ten years? What will be different if we reconvene here five years from now? How will the world change?
领英推荐
The first thing hot on everyone's mind is regulation and controls. All of us who work in this space are thinking about a lot, but I will be guided by historical precedent. Let's pick an example: aviation. Planes were built, accidents happened, and then we realized the need for air traffic controllers and traffic lights. Similar situations occurred with automobiles and nuclear power.
People understood the potential dangers and decided to establish regulations. The same pattern can be seen in the AI space. European countries have taken the lead in implementing regulations, while the US may adopt a different framework. Eventually, an international organization might be formed to set global standards.
The challenge lies in the rapid pace of change in the digital world. Government agencies may take years to understand new issues and develop policies, by which time we are already many generations ahead. However, imposing a ban for a short period may not be the solution. Some calls for moratoriums may come from those who failed to catch up with the advancements and seek time to recover. Industry-led regulations have been successful in other sectors, so similar conversations within the AI industry might be necessary.
There are friction points between human and digital time, as well as future time. AI systems can process information and generate solutions within seconds, while human thinking operates on a circadian rhythm. This intersection between AI's rapid thinking and human contemplation is an intriguing aspect to observe.
Looking into the future, I see three time frames. In the immediate short term, intermediaries will emerge to guide enterprises in adopting AI. In the medium term, more government regulations and considerations of equity will arise. AI's role in addressing global issues, not just limited to the developed world, will be explored. In the long term, there may be advancements in AI-enhanced humans, such as neural link technologies and brain augmentation. Utilizing the untapped potential of the human brain, combined with AI, could lead to even greater achievements.
Elaborating further:
In the immediate term, I see the rise of intermediaries who will create context and act as consultants or prompt architects to support enterprises entering the AI space.
Large enterprises will adopt AI in the medium term, accompanied by government regulations, traffic lights, and guardrails.
We should also consider questions of equity and AI's role in addressing basic human problems worldwide, not just in the first world. AI could potentially play a role at the bottom of the pyramid as well.
Looking into the long term, we're already witnessing advancements in AI-enhanced humans, such as neural links and implants that assist individuals with physical limitations.
However, there is untapped potential in the human brain. New technologies may emerge to unlock this potential in conjunction with AI, allowing us to accomplish even greater feats. Imagine harnessing the idle processing power of our brains and using it to augment AI for various tasks—it would be phenomenal. A Human Intelligence Augmented Artificial Intelligence of sorts.
This concept could be likened to the Napster of human brains in a peer-to-peer network, although the exact form it would take is uncertain. Humans have previously attempted to utilize idle computing power for projects like the search for extraterrestrial intelligence. Similarly, we could use AI to better understand our own brains and potentially address human problems.
This journey has just begun, and there is much to be excited about. It's not something to fear, but rather a source of enthusiasm and discovery. Of course, checks and balances are necessary to ensure responsible development, as humans have done for generations.
POSTSCRIPT
Exciting (and scary? or not?) stuff indeed. Stay Tuned. Follow the hashtag #DEEPakAI on LinkedIn to follow all my AI related content/posts.