PREDICTIVE

Machine Intelligence in National and Cyber Security

Dr. Launchbury, Director of the Information Innovation Office, DARPA

 

Thinking about machine intelligence as a sentient and potentially malicious entity is outdated. Instead, today, it is more appropriate to think of it as software that has some facility for independent action, analysis, and decision making. To the point where it can even reprogram itself to tackle new and unexpected challenges, even. In this session, Dr. John Launchbury, Director of the Information Innovation Office at DARPA, speaks with Eric Newcomer, a reporter covering global technology with Bloomberg.

Highlights:

  1. The easiest problems for machine intelligence to attack and information processing problems. Truck drivers are safe because you have to move through space, through the real world.
  2. DARPA was created in reaction to the Sputnik revelation and intends to keep the United States competitive technologically by stating, “I think it is possible to do this” when confronted with a challenge
  3. Development of AI comes in waves. First wave machine intelligence works within rules and structures it is programmed to work with. Second wave AI works on its own to understand a problem and develop a solution. Third wave machine intelligence will finally be in a position to begin displacing human workers in tasks that are somewhat predictable or can rely on educated decision making, such as matching cancer patients to treatments.

Machine Intelligence in National and Cyber Security

What about you?

Do you believe your job, or parts of your job, can be handled by machine intelligence, or do you believe it is secure?

Leave a Reply