Independent Student Newspaper Since 1969

The Badger Herald

Independent Student Newspaper Since 1969

The Badger Herald

Independent Student Newspaper Since 1969

The Badger Herald

Advertisements
Advertisements

World is safe from Artificial Intelligence, Ken Jennings is not

Artificial intelligence appears to be the last frontier of computing technology. Enabling a machine to answer open-ended questions accurately has been quite an elusive task, but the programmers at IBM seemed to have cracked the riddle. Of course, as quickly as the AI topic pops up, so too does the uni-brow of the sci-fi geek compulsively fantasizing about cyber-organism domination. Although this may be a legitimate concern for many, the IBM program nicknamed “Watson” does not appear to be conceiving a Terminator-like plan for human annihilation. Instead, Watson has targeted Alex Trebek’s game show. Of course, if Watson were to go haywire, I’m sure we could all rely on Schwarzenegger and Christian Bale for salvation. In reality though, I think it’s safe to allow AI to conquer the realm of trivia, and that is exactly what IBM’s directors did.

Earlier this week, ABC network pitted man against machine in an IBM “Jeopardy” Challenge. In an effort to demonstrate its technological prowess, IBM developers showcased Watson, their deep “thinking” question-and-answer computer program. Watson faced off against two of Jeopardy’s most celebrated human programs: Ken Jennings, who in 2004 went on a 74 game rampage earning him $3,022,700, and Brad Rutter, the most winning Jeopardy contestant of all-time. Rutter accumulated $3,255,102 in assets and a pair of Chevy Camaros – fair competition to say the least.

Reminiscent of chess grandmaster Gary Kasparov’s 1997 defeat by IBM’s computing system Deep Blue, the American multinational has yet again conquered the cerebrums of arguably the world’s biggest nerds. Open-domain questioning is a much more intricate process than a formalized game of chess, though. The latter relies on advanced algorithms to make the best possible play at the appropriate time – Watson relies on statistics from language, wealth of Internet information and algorithmic techniques to put forth an answer. It performs 33 billion operations a second to derive the statistics needed to pass its confidence threshold prior to buzzing in. Thus, cracking Jeopardy was a much more ambitious accomplishment.

Advertisements

Consequently, the Watson program augurs yet another revolutionary step: Computer programs are mastering the task of information aggregation. The success of the deep question-and-answer format underlines a current challenge that we face in our quest for information. Entering keywords into a Google search engine results in a list of documents that are ranked by popularity rather than relevance. The user must then sort through and decipher the results. Watson’s beauty, however, lies in its ability to precisely identify what exactly is being asked and zero in on the most relevant answer – all in a much timelier fashion.

Think of a Yahoo! Answers website that nails it every time.

Questions such as, “How many people were mauled to death by bears last year in North America”? (28) or “How long can a cockroach live with its head cut off”? (several weeks) would be answered in well under a second with exact precision. The implications this holds for all fields, including health care and business investment, are endless, and not even a perfect Watson would be able to forecast that number.

We must not get ahead of ourselves just yet. IBM’s program has taken 20 researchers more than three years to create, and they have spent millions of dollars polishing Watson so that he is presentable to the public.  Although Watson handily beat two of our arguably sharpest minds, there remain many kinks to work out. Artificial Intelligence is, on the whole, still in a very nascent stage. On day two of the tournament, Watson quite unintelligently answered, “What is Toronto”? for the final Jeopardy clue: “Its largest airport is named for a World War II hero; its second largest for a World War II battle.” The category was U.S. cities. This, and the fact that Jeopardy organizers agreed to omit audio-visual categories and clues that required Trebek’s instruction reveals Watson’s limitations.

I think it is safe to say for now that we will not face extermination at the hands of cyborgs or Watson; if we do, let’s just hope Watson will miscalculate again and release the hounds of James Cameron on Canada.

Dennis O’Reilly ([email protected]) is a senior majoring in economics.

Advertisements
Leave a Comment
Donate to The Badger Herald

Your donation will support the student journalists of University of Wisconsin-Madison. Your contribution will allow us to purchase equipment and cover our annual website hosting costs.

More to Discover
Donate to The Badger Herald

Comments (0)

All The Badger Herald Picks Reader Picks Sort: Newest

Your email address will not be published. Required fields are marked *