GenAI-Driven Attacks Require Changes to Application and Data Security Practices and User Monitoring End-user spending on security and risk management (SRM) in India is forecast to total $2.9 billion in 2024, an increase of 12.4% from 2023, according to a new forecast from Gartner, Inc. Indian organizations will continue to increase their security spending through 2024 due to legacy IT modernization using cloud technology, industry demand for digital platforms, updated regulatory environment, and continuous remote/hybrid work. “In 2024, chief information and security officers (CISOs) in India will prioritize their spending on SRM to improve organizational resilience and compliance,” said Shailendra Upadhyay , Sr Principal at Gartner. “With the introduction of stringent government measures mandating security breach reporting and digital data protection , CISOs are facing heightened responsibility in safeguarding critical assets against evolving cyber threats.” Gartner a...
As a kid I loved Star Trek. Not the spaceship stuff, but the interactive computer systems. The ability to query a computer system with your voice and get an intelligible response is probably the final frontier of computing as we know it. Of course, the underlying artificial intelligence (AI) to interpret and execute a response is part of the whole process. I think this is exactly where technology conglomerates are now focusing their energy.
Already the latest generation of smartphones support both voice recognition and voice playback, technology which has jumped significantly in recent years. Gone are the days of having to train your PC how to understand you by repeating written excepts continuously to improve accuracy. Furthermore, the robot voices have improved drastically. The only thing which seems to be suffering is the AI part in the middle.
Both Apple's SIRI and the experimental Android IRIS smartphone AI rely on an Internet connection to answer even the most basic of commands. The approach is to offload all the intelligence into the network.
This means the distributed processing capacity of the back-end servers would be much faster at interpreting commands compared to the processor of a phone. However, I feel the future will be to bring a degree of processing back into the device itself. For example, from a programming point of view, it doesn't require any AI for a smartphone to respond locally providing an answer to a query for the time in another city.
The difficulty is processing the speech samples and extracting the key words for a probable response. AI systems in general have had a troubled history. Some will argue a chess program is a good example of AI such as IBM's Deep Blue, which beat chess champion Garry Kasparov in 1997.
However, much of the credit of AI systems has been relegated to everyday functionality, such as the weighted search drop probability algorithms used by Google and other search engine providers. Millions of dollars have been spent over the decades in the quest to develop a mechanical equivalent of the human brain, a journey which has proved illusive. I believe the search engine giants hold the answer to what we would consider true AI.
If you type a sentence into Google, the first thing that happens is the meaningless filler words like “if, or, but” are dropped. Only the adjectives are kept. Since most people use the same phrases in the same language the probable search response that has worked for others in previous searches is displayed as a solution. Moving this concept forward to another level seems to be the natural progression of technology.
However, it would seem that for now the hardware is still the limiting factor in a truly interactive AI solution. Any smartphone solution would require a reasonable Internet connection and a considerable amount of back-end processing. Moving this processing power into your phone will be the next challenge, since chances are a data connection will not always be available.
As CPU technology becomes yet smaller and faster, and memory larger and cheaper, I don't think it will be too long before you will be able to stumble out of a bar at 2am and with a slurred expression say “call me a taxi, I want to go home.” The phone will trace your location via a GPS or cell location, and it will either order the taxi electronically or speak to a dispatcher informing the smartphone's owner of the taxi's ETA.
The days of punching away on your mobile is certainly numbered. Imagine activating the personal assistant (PA) and saying something like “send an email to Bob and Alice, thanking them for a wonderful evening.” The phone will interpret this, correlate the past week of address locations from the phone's memory, find a match for Bob and Alice in the address book, create a suitable thank you note and send it.
The PA should be adaptive and learn it's owner's habits and movements. All of this behaviour is only possible with faster silicon and larger memory, but if you ask me, AI is definitely the next killer app.
By Craig Sutherland
Comments
Post a Comment