Another article hot off the press extolling the virtues of Edge AI.
Although the author is CEO of Xperi, a company that has a competing (although apparently not truly neuromorphic) Edge AI processor (
https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-289368), it is nevertheless good publicity for the Edge AI market in general.
Beyond taking legal boilerplates seriously, however, a meaningful conversation is emerging around the implications of applying AI and machine learning.
www.forbes.com
No Room For Privacy Complacency In The Era Of Cloud-Based AI
Jon Kirchner
Forbes Councils Member
Forbes Technology Council
COUNCIL POST| Membership (Fee-Based)
Jun 23, 2023,08:15am EDT
Jon Kirchner, CEO,
Xperi.
GETTY
While consumers worldwide have largely ignored invitations to review lengthy data collection policy disclosures before accessing the services and experiences they want to enjoy, advocates for protecting personal digital information are wondering out loud if the real-world implications of cloud-based artificial intelligence (AI) will finally knock consumers out of their "privacy complacency."
Beyond taking legal boilerplates seriously, however, a meaningful conversation is emerging around the implications of applying AI and machine learning (ML) to gather, analyze and process information in the cloud compared to edge computing environments—in which processing and data storage takes place closer to where data is created or consumed. It turns out that integrating AI and edge computing opens the door to significant benefits for those interested in enhancing data-powered experiences while protecting their personal information.
Cloud-based AI/ML services have crossed an important line when it comes to personal digital information. The technology no longer just stores the data. It specifically gathers, optimizes, trains on and monetizes data from virtually every public online source—including cloud services and connected devices. The early use cases have demonstrated that this technology can—and will—indiscriminately access and absorb contextual information from any digital source to do its job.
This is especially concerning as elements in the most sensitive sectors of society—including healthcare—rush to integrate cloud-based AI tools into their offerings. Some online mental and emotional support services, for instance, have controversially integrated AI chatbot technology into therapy and counseling sessions. On at least one occasion, cloud-based AI was deployed without the informed consent of patients/clients.
An Irrational Bias For Cloud Impedes Privacy-By-Design
In the meantime, terabytes of data are being "hoovered" into a small handful of cloud service providers (CSPs) as users engage cloud-based AI applications from mobile devices, smart homes, connected cars—and even biologically embedded technologies, such as connected pacemakers. It represents a colossal concentration of data that should raise serious alarms and prompt consumers—as well as business leaders across all industries—to develop robust privacy-by-design frameworks.
Originally developed to prevent personal data breaches by incorporating privacy considerations into the architecture of digital offerings, privacy-by-design also seeks to ensure that consumers retain control over their information. It is an objective that is difficult to achieve if all data manipulated by AI technologies ends up in cloud environments.
There can be little doubt about the positive role that cloud computing has played in democratizing access to processing and storage capacity. At a business level, cloud computing has allowed small- and mid-sized organizations to take advantage of technology-enabled operations that were once the exclusive domain of the Global 2000.
It has also opened the door for corporate behemoths to become more agile, replicating the market responsiveness once associated with smaller entities. From a consumer point of view, cloud and mobile apps have introduced a wide array of new options that have greatly expanded the experiences available to individuals.
A good argument, however, is emerging for a more balanced approach to the future of the global digital economy. As a society, we should engage in a more robust discussion about how we can get the most out of AI/ML to improve end-users lives without aggregating every citizen's personal information into a handful of CSPs. Indeed, compelling technical benefits are associated with constructing a global information infrastructure that keeps the bulk of our information where most of it belongs—locally and in the control of those who truly own the data.
Integrating The Power of AI With Edge Computing
For instance, edge computing and AI can make our roadways safer by enabling connected car technologies—including in-cabin cameras and sensors—that notice when a driver is getting drowsy. According to the most recent Center for Disease Control (CDC) survey, approximately one in 25 adult drivers (18 years or older) reported falling asleep while driving. The CDC estimates this results in over 6,000 fatal crashes and 50,000 injuries yearly.
This challenge offers an excellent example of a use case in which data processing must occur at the edge because concepts like "latency" matter. It is difficult to see how drivers, passengers and pedestrians can be kept safe if data must be sent to a distant cloud resource for AI analysis.
There is also the issue of adoption. One can safely assume that many, if not most, consumers will be reluctant to enable functions that collect sensitive behavioral data if it is sent to the cloud for AI processing. (At a more practical level, it is probably not a good idea to deploy vitally important connected car solutions that only work when there is a cell signal—a requirement for remote access to cloud resources.)
There is simply no reason to move all information out of its local environment all of the time. There are growing use cases for edge computing technologies to be applied—and controlled—by consumers to improve accuracy and performance and protect privacy.
How?
Cloud-based AI needs the brute force of big data analytics before arriving at a conclusion or result. Edge computing—done right—applies machine learning inference from the opposite direction. The most innovative edge-based computing technologies use a small amount of real-time information to derive an accurate result. It is like the old game "Name That Tune."
In short, consumers and enterprises interested in controlling the management of personal information should encourage providers of digital products and services to calibrate how and when data is used in cloud-based AI environments.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives.
Do I qualify?
Follow me on
LinkedIn. Check out my
website.
. But I genuinely believe that AI will take over Humans someday. But that’s for another day!
region!) and discovered the following Spring/Sommer 2023 hospital newsletter with more info on the said lung nodule programme: 
this post about a newly launched “lung nodule programme” by an MD at a California hospital on a six-month demo trial with a state-of-the-art robotic-assisted bronchoscopy system, raising funds for “an Artificial Intelligence System to detect signs of cancerous lung nodules up to a year earlier than manual-only review of x-rays” without Brainchip being involved?!
www.designnews.com
With the remaining 12.5%, I have a pain ratio of 7:1, which is bearable.
