A roster of officials from government, academia and industry gathered in Munich Germany at a Security and Cyber Security Conference to discuss how future workforces must marry the power of artificial intelligence with expertise only a human can provide. “Looking at the next generation of national security professionals, I want policy people who can code and coders who can do policy,” said the former head of the National Security Agency (NSA) General Paul Nakasone at the Munich Cyber Security Conference. “Five years ago Baby Boomers were replaced by Gen Z’ers, and five years from now it’ll be people born in 1997 and it’s a workforce that understands data, large language models and speaks a lot of languages, including computer languages.”
Nakasone, who is now the director of Vanderbilt University’s Institute of National Security, says that is precisely what he started to see when he was leading the US’s largest spy agency, young recruits were fundamentally wired differently than the operators and analysts who came before them. “The side that is going to win is the side that can integrate AI as quickly as possible in what they do, and speed matters here,” he said. “And where does that come from? From people who think differently” who can make AI a force multiplier, not a replacement.[1]
Nakasone was speaking on a panel that was trying to imagine the AI workforce of the future. Nakasone said while he was leading the NSA and US Cyber Command, there were things operators did that AI couldn’t replicate. He recalled operators who had, for example, been listening to an adversary for years and could pick out a subtle change in their tone, a very human ability. “There is something different in that voice,” he said. “They could tell something is different” and that helped the team make better decisions. AI will block and tackle, he said, but humans will still be needed to think through ethical and moral implications.
Human neurodiversity as a superpower - Peter Kant is the chairman and CEO of a Virginia-based company called Enabled Intelligence, which, he told the Munich audience, provides the human part of AI, mostly for defense and intelligence applications. As he sees it, the way to make AI useful, reliable and accessible to humans is to evaluate AI with the broadest neurodiversity possible in order to ensure that the answers the AI comes up with actually make sense. “Most of our team is looking at how to find hallucinations,” Kant said. “Now you can get something that looks really good that has fake legal references and fake patents… [or] emails don’t sound so bad and it requires a different level of education” to see the flaws.
AI is only as good as the data it learns from. Large language models, or LLMs, like OpenAI’s GPT, will produce flawed or biased results if the datasets it trains on are flawed and Kant says he’s been working with neurodiverse individuals who are particularly good at spotting and eliminating those problems.
Humans really matter. AI automates the mundane so humans can find ways to make opportunities happen. Which is why, Kant told the audience, he has a neurodiverse workforce. The unique aptitude of some neurodiverse individuals for detail orientation, pattern recognition and focus during seemingly repetitive analysis, Kant said, has been proven to help clean up the data used to develop AI models for defense and it has proven to be much less susceptible to typical human biases.
Kant said their neurodiversity becomes a superpower as they, for example, assess millions of satellite images and detect objects other people miss, everything from camouflaged missile launchers to unexploded ordnances in a Ukrainian field. “Humans do critical thinking better than anything else we know of,” he said.
This article is shared at no charge for educational and informational purposes only.
Red Sky Alliance is a Cyber Threat Analysis and Intelligence Service organization. We provide indicators of compromise information via a notification service (RedXray) or an analysis service (CTAC). For questions, comments or assistance, please contact the office directly at 1-844-492-7225, or feedback@redskyalliance.com
Weekly Cyber Intelligence Briefings:
- Reporting: https://www.redskyalliance.org/
- Website: https://www.redskyalliance.com/
- LinkedIn: https://www.linkedin.com/company/64265941
Weekly Cyber Intelligence Briefings:
REDSHORTS - Weekly Cyber Intelligence Briefings
https://register.gotowebinar.com/register/5207428251321676122
[1] https://therecord.media/putting-the-human-back-into-ai-is-key-nakasone/
Comments