3 important generation AI issues that still need to be resolved
The adoption of generative AI in technology is potentially greater than when the internet was introduced. It disrupts most creative endeavors and isn’t performing as well as it will be by the end of the decade.
Gen AI will force us to rethink how we communicate, collaborate, create, problem solve, govern, and even how and if we travel – and that’s far from an exhaustive list. I expect that once this technology matures, the list of things that haven’t changed will be much shorter than the list of things that have.
This week, I’d like to focus on three things we should start discussing that represent some of the biggest risks of generative AI. I’m not against technology, and I’m not stupid enough to suggest that it be paused because it would be impossible to pause it now.
What I’m suggesting is that we start looking at mitigating these issues before they cause significant damage. The three issues are data center load, security, and relational damage.
We’ll end with my Product of the Week, which might be the best upcoming electric SUV on the market. I’m suddenly looking for a new electric car, but I’ll talk about that later.
Data Center Loading
Regardless of all the hype, few people are still using generative AI, let alone using it to its full potential. The technology is processor- and data-intensive while being very person-centric. It will therefore not be possible to have it reside solely in the cloud, mainly because the resulting size, cost and latency would be unsustainable.
Just as we’ve done with other data- and performance-driven apps, the best approach will likely be a hybrid where the processing power is kept close to the user. Nevertheless, massive data, which will require aggressive updating, will need to be loaded and accessed more centrally to protect the limited storage capacities of client devices, smartphones and PCs.
But, because we are talking about an increasingly intelligent system that sometimes, like when used for games, translation or conversations, will require very low latency. How the load is distributed without hurting performance will likely determine whether a particular implementation is successful.
Achieving low latency will not be easy because even though wireless technology has improved, it can still be unreliable due to weather conditions, tower or user location, maintenance outages, man-made or natural disasters and less than complete global coverage. AI must work both online and offline while limiting data traffic and avoiding catastrophic outages.
Even if we could centralize all of this, the cost would be excessive, even though we have underutilized performance in our personal devices that could mitigate much of this expense. Qualcomm is one of the first companies to flag this as an issue and is working hard to fix it. Still, expect it to be too little too late, given how quickly generative AI is advancing and how slowly technology like this is being developed and brought to market.
I was an internal auditor specializing in security and a competitive analyst trained in legal ways to penetrate security. I learned that if someone can get enough data, they can more accurately estimate the data they don’t have access to.
For example, if you know the average number of cars in a company parking lot, you can, with reasonable accuracy, estimate the number of employees in a company. You can usually scour social media and determine the interests of key company employees, and you can look at job postings to determine the types of future products the company is likely to develop.
These large language models collect huge amounts of data, and I expect that many of the things analyzed by these LLMs are or should be confidential. Also, if enough information is collected, the gaps resulting from what is not scanned will be more and more derivable.
This scenario does not only apply to corporate information. With the type of personal information readily available, we will also be able to learn much more about users’ privacy.
Employers will be able to more accurately locate whistleblowers, disgruntled or disloyal employees, employee misbehavior, and employees who illicitly take advantage of the company. Protection against a hostile entity deriving confidential information about you, your business, or even your government becomes more viable with far greater accuracy than I have enjoyed as an auditor or competitive analyst.
The best defense is likely to create enough misinformation that the tools don’t know what’s real and what’s not. However, going this route will also make connected AI systems much less reliable overall, which would be nice if only the competitor was using those systems. However, it is likely to compromise the systems the company wishing the protection could use, leading to an increasing number of bad decisions.
Companies like Mindverse with its MindOS and Suki with its employee complementing avatars are pitching the future personal use of generative AI as a tool that can present itself as if it were you. As we use tools like this, our ability to determine what is real and what is digital will be greatly reduced, and our opinions of the people who use these tools will reflect the tool more than the person. .
Imagine your digital twin doing a virtual interview, either being the face of your presence on a dating app or taking over for much of your daily virtual interactions. The tool will try to respond to the person interacting with it, it will never be tired or grumpy, and it will be trained to present you in the best possible light. However, as he goes down this path, he will become less and less like who you really are – and will probably become much more interesting, attractive and more equal than you ever could be.
This will cause problems because, much like actors dating someone who has fallen in love with a character the actor once played, reality will create later breakups and loss of trust.
The easiest solution would be to either learn to behave like your avatar or use them for interactions with friends and colleagues. I doubt we will either, but those are the two most viable approaches to mitigating this coming problem.
Generative AI is amazing and will dramatically improve performance as it takes hold in the market and users reach critical mass. Still, there are significant issues that will need to be addressed, including excessive data center loading, which should lead to hybrid solutions in the future, the inability to prevent secrets from being derived from these huge language and a considerable reduction in interpersonal trust.
Understanding these future risks should help avoid them. However, the fixes aren’t great, suggesting that we’ll likely regret some of the unintended consequences of using this technology.
The Fisker Ocean
My Jaguar I-Pace bit the dust last month due to a towing accident that damaged its battery. The result was a $100,000 estimate to fix the car, which is now worth closer to $40,000. I expect USAA, my insurer, to total the car. So, I searched for replacement electric cars, and overall the availability sucks.
I’m likely to get another Jaguar I-Pace mainly because I don’t want to wait months or years for my own car again. Currently I am sharing my wife’s Volvo XC-60 and am having a lot of scheduling issues where we both need the car at the same time. I went shopping for a new electric SUV, and the best I found was the Fisker Ocean.
The all-electric Fisker Ocean (Image credit: Fisker)
As with most electrical appliances, the wait for a new one is months, and I can’t afford that wait. Of the electric cars available this year, the Fisker Ocean hit all the boxes. Its features include:
- 350 mile range (bar is 300 miles)
- Your smartphone can be the key to your vehicle
- Reverse charging, so your car can power your home during a power outage
- An impressive 0-60 time of around 3.6 seconds (I like the performance)
- A roof of solar panels to increase range and provide backup power
- A convertible-like mode (which really opens up the car)
- One of the cleanest designs on the market.
The Fisker Ocean is an impressive car. If I could wait until the end of the year to get one, I would order it in a moment. Unfortunately, it is not the case. Nevertheless, the Fisker Ocean is still my product of the week.
The opinions expressed in this article are those of the author and do not necessarily reflect the views of ECT News Network.
Leave a Reply