Introduction
This week we journey into the world of Artificial Intelligence. The term Artificial Intelligence itself has largely become the generic term used to describe a variety of related but different concepts, often wrapping together AI, Machine Learning and Generative AI as if they all relate to the same thing, or as if there is a single technology that is, in fact, AI. These related but different technologies differ in scope, function, techniques used, and they also differ in terms of trustworthiness and reliability. But generally speaking, Artificial Intelligence focuses on techniques and applications that create the appearance of human cognition in or from a machine. Machine learning focuses on the development of algorithms and statistical models that enable computers to perform tasks without being explicitly programmed for those tasks. And Generative AI focuses on creating new data, such as images, music, and writings, that mimic the patterns presented in the training data.
Technical Issues in AI
Much of what we will discuss here tackles the technical issues that face the industry and innovators in this space. Specifically, we will discuss the various practical manifestations of Artificial Intelligence, ML, and Generative AI as they currently exist in 2024, and are likely to exist in the near future. We will attempt to separate fact from fiction with respect to what these technologies can currently do, what the technology does well versus what the technology struggles with, and what the industry can expect moving forward.
Challenges and Reliability
We will also discuss the thorny problem of hallucinations, the impact of Artificial Intelligence learning from its own creations, the steps that are being taken to ensure reliability and accuracy, and how to verify that AI is safe, secure and trustworthy.
Expert Insights
First, we will hear from Jason Alan Snyder, who is Global Chief Technology Officer at Momentum Worldwide, which is a part of the Interpublic Group. Jason is a futurist, technologist, inventor, author and lecturer on topics relating to the metaverse, artificial intelligence, machine learning and Cognitive Robotic Processes Automation. Jason, a long-time friend, has been recognized as one of the top 30 experts in the world on topics dealing with the metaverse, virtual reality, augmented reality and blockchain.
Asked to give his preliminary thoughts on where AI technology stands today, Snyder explained that AI certainly does not have agency, and it is not going to take over the world tomorrow. This prompted me to follow up, because saying that AI will not take over the world tomorrow implies that AI could, at some point, actually take over the world, which Snyder responded “absolutely, that is a possibility.” Snyder also spoke about how AI has no agency and does not have the same shared, human context, which he calls intersubjectivity.
“It has zero context. It has no idea what Adolf Hitler means. It has no idea what Elon Musk means,” Snyder explained. “That means something to us, but not to the platform, and we keep forgetting that every time we sit down in front of it because it feels like it has agency. It has no agency.”
Security Concerns
We also hear from Dr. Malek Ben Salem, who is Managing Director and a Data & AI lead at Accenture Security, where she focuses on security for emerging technologies, including quantum, blockchain, AI and the Metaverse. Among other things, Ben Salem explained how memory limitations cause AI to break down and leak sensitive information and technology.
“It’s called a divergence attack,” Ben Salem said. “You keep interacting with the AI until it just loses the context and has to completely switch to another context and starts generating information. And by way of doing that, it runs the risk of leaking sensitive information. I mean, obviously there are memory limitations, right, as you’re interacting with the AI. At some point you’re going to hit that memory limit and then it will lose track of the context because it hit that memory limit. So you can increase the memory limit, but, you know, no matter what, at some point you’re going to hit it. So I think you’re always going to find a way to attack the AI.”
Conclusion
To hear this entire conversation, listen wherever you get your podcasts (links here). Or visit IPWatchdog Unleashed on Buzzsprout.
Gene Quinn Gene Quinn is a patent attorney and a leading commentator on patent law and innovation policy. Mr. Quinn has twice been named one of the top 50 most influential people […see more]
Warning & Disclaimer: The pages, articles and comments on IPWatchdog.com do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of IPWatchdog.com.
- 0 Comments
- Ai Process
- Artificial Intelligence