By Byron V. Acohido

San Francisco, Calif. — The amazing digital services we have today wouldn’t have come to fruition without the leading technology and telecom giants investing heavily in R&D.

Related: GenAi empowers business

I had the chance to attend NTT Research’s Upgrade Reality 2024 conference here  last week to get a glimpse at some of what’s coming next.

My big takeaway: GenAI is hyper-accelerating advancements in upcoming digital systems – and current ones too. This is about to become very apparent as the software tools and services we’re familiar with become GenAI-enabled in the weeks and months ahead.

And by the same token, GenAI, or more specifically Large Language Model (LLM,) has added a turbo boost to the pet projects that R&D teams across the technology and telecom sectors have in the works.

The ramifications are staggering. The ability for any human to extract value from a large cache of data – using conversational language opens up a whole new universe of possibilities.

The power of conversations

One small example is a souped-up Jibo smart home assistant — a prototype — that can do much more than lock the doors, turn out the lights and set the thermostat. Thanks to GenAI, users can engage this prototype in conversations that get steadily richer over time.

Heidbrink

NTT Research is testing its Jibo protype as a chatty, mindful digital companion oriented to assisting the elderly in multifaceted ways. Sensors scattered around a home keep track of motion, temperature, CO2 levels, light levels and sound. A baseline gets established, deviations get analyzed and responses automatically get fine-tuned.

This all gets done leveraging well-established AI algorithms — but  GenAI takes it to another level, says Chris Heidbrink, NTT Research senior vice president of AI & Innovation.

By factoring in human language cues, Jibo over time can start to detect sentiment and potentially identify health conditions based on conversations. “What we’re doing is combining traditional AI with quality data —  and then bringing in GenAI is like adding polish to it,” Heidbrink  told me. “GenAI allows us to plug in many different things, combine them together and have really deep conversations about them.”

Tech giants out front

Jibo is a microcosm of how GenAI is turbo boosting R&D prototypes of all kinds. Meanwhile, the dust storm clouding the tech horizon is being kicked up by enterprises in all sectors racing to deploy GenAI in support of their entrenched business models.

This GenAI gold rush is being led by the marquee tech giants. Like me, you may be beta testing Adobe’s “Ai Assistant” prototype for Acrobat that allows you to type conversational commands directly into PDF documents. On my SEA to SFO flight, I sat next to a Meta software engineer and we chatted about how Microsoft’s $10 billion  investment in OpenAI/ChatGPT is all about integrating ChatGPT into Windows and Office, while Google’s Gemini services is all about infusing GenAI into Google Search, Google Docs and YouTube.

Likewise, Facebook LLaMA is Meta’s attempt to extract more value from its core asset, Facebook users’ digital footprints. This, of course, raises profound privacy and cybersecurity questions that are just starting to heat up with the rising tide of GenAI-infused deep fake attacks.

Cybersecurity conundrum

Somewhat ironically, the cybersecurity industry itself is scrambling to integrate LLM into emerging security platforms and frameworks to mitigate deep fakes, as well as to get in a better position to address sure-to-come iterations of cyber attacks enhanced by GenAi. (Stay tuned for Last Watchdog’s RSAC Insights podcasts from RSAC 2024, just around the corner.)

I broached this topic at Upgrade Reality 2024 with Moshe Karako, CTO of NTT Innovation Laboratory in Israel. On a whim, while waiting for a flight to Tokyo, Karako was able to persuade Microsoft’s Copilot chat tool to violate Microsoft policy and solve a  captcha to gain him access to a secured website page.

Karako

Moshe used tried-and-true social engineering tactics, such as misspelling words and using persuasive language, to lower Copilot’s guard and manipulate the conversation in his favor. “All it took was playing with prompts to convince it to do what I needed,” Karako says. “And there’s no active solution today that can prevent this.”

Here we go again. Remember how email spam, evolved into phishing attacks, ransomware and advanced persistent threats? This evolved over the past 20 years as business networks advanced from on-premises data centers to hybrid cloud. Along the way, cyber exposures mushroomed. Now GenAI has set us up for a repeat of that cycle — only at a breakneck pace of change.

The hype over the impact of GenAI is just getting started. I heard Vab Goel, founding partner of NTTVC, declare that GenAI will trigger 100X change 100 times faster the we saw in the Internet revolution. Another executive, Rajeev Shah, founder and CEO of Celona.io, I thought, put it best. Speaking on a panel discussion about the transformative potential of GenAI, Shah said this:

Shah

“Actually, I think, as a Silicon Valley (company) founder. It’s very rare, and it’s actually the first time in my entire career, that I have been confronted with a technology that neither can I fully understand, nor can I fully grasp the potential. I don’t think any of us have fully internalized yet how transformative AI can be.”

So where will this democratization of AI take us over the few months and in next couple of years? That’s the breakneck speed revolution we’re all about to experience. I’ll keep watch and keep reporting.

Acohido

Pulitzer Prize-winning business journalist Byron V. Acohido is dedicated to fostering public awareness about how to make the Internet as private and secure as it ought to be.

April 15th, 2024