Skip to main content

I’ll Take Supercomputers for $2,000, Alex

I

IBM has famously developed systems that have taken on the grandmasters of chess—and given them a pretty serious run for their money, defeating world champ Gary Kasparov back in 1997. Now, IBM is going after the big bucks with Watson, a new question-answering system that will compete on long-running U.S. television quiz show Jeopardy. Watson will not only have to understand the questions put to it in real time on the show against human contestants, but it will also have to “know what it knows,” determine its level of confidence in being able to answer a question, and buzz in—just like a real human. But Watson is more than a publicity stunt from IBM: it represents an effort to elevate computer intelligence and human-computer communication to a new level—and that’s technology that can be applied to much more than television quiz shows.

“The essence of making decisions is recognizing patterns in vast amounts of data, sorting through choices and options, and responding quickly and accurately,” said IBM head Samuel J. Palmisano, in a statement. “Watson is a compelling example of how the planet—companies, industries, cities—is becoming smarter.”

Watson uses a parallel analytical architecture, but will not be connected to the Internet or other external data sources during the competition: it will have to work solely with the knowledge base pre-configured by IBM’s engineers. Watson will also have to be a particularly adept reader: Jeopardy questions are well known for their wordplay, irony, puns, and humor—things humans enjoy, but which have been known to send computers off the deep end.

“The challenge is to build a system that, unlike systems before it, can rival the human mind’s ability to determine precise answers to natural language questions and to compute accurate confidences in the answers,” said Watson product leader Dr. David Ferrucci, in a statement. “This confidence processing ability is key. It greatly distinguishes the IBM approach from conventional search, and is critical to implementing useful business applications of Question Answering.”

There’s no word yet on when Jeopardy shows featuring Watson will be taped or aired…but we bet the Jeopardy writing staff is looking forward to the challenge.

Geoff Duncan
Former Digital Trends Contributor
Geoff Duncan writes, programs, edits, plays music, and delights in making software misbehave. He's probably the only member…
A dangerous new jailbreak for AI chatbots was just discovered
the side of a Microsoft building

Microsoft has released more details about a troubling new generative AI jailbreak technique it has discovered, called "Skeleton Key." Using this prompt injection method, malicious users can effectively bypass a chatbot's safety guardrails, the security features that keeps ChatGPT from going full Taye.

Skeleton Key is an example of a prompt injection or prompt engineering attack. It's a multi-turn strategy designed to essentially convince an AI model to ignore its ingrained safety guardrails, "[causing] the system to violate its operators’ policies, make decisions unduly influenced by a user, or execute malicious instructions," Mark Russinovich, CTO of Microsoft Azure, wrote in the announcement.

Read more