/ps: 105 crid: 10132:1-allec-ci-yottaa-1 cc:de
By now most of you know that the last US presidential election had more to do with Russia and Facebook than pure voter preferences. Many of those who might have been inclined to vote for Clinton were discouraged from voting. Trump was cast in a more attractive light, and those who were impressed by him were motivated to vote.
The biggest irony is that Trump has been outspoken against fake news, even though he likely is one of the biggest single sources of it. Now don’t get me wrong, I’m kind of glad he is president, because unless this fake news problem is fixed, the U.S. could be on a path toward eliminating democracy altogether and eventually becoming a Russian or Chinese State.
Very often when the president speaks, many of us in tech are reminded that we need a universal unbiased tool to help us identify and disregard false information before we make a bad decision based on it.
IBM just showcased Project Debater, and it just won against a top human debater using actual facts. We need the production version of this tool yesterday.
I’ll explain and then close with my product of the week: the Plantronics Voyager 6200 UC, which I’ve found is the best headphone set for traveling.
/ps: 55 crid: 10027:a-nmg_tnw_160-1 cc:de
The Problem With Fake News
The problem with fake news is that it is getting worse over time. We have always had issues with politicians and con artists trying to convince us down is up and to get us to do things that, were we informed, we’d never do.
I now get what seems like at least one daily phone call from someone pretending to be from the IRS, U.S. Treasury (just got one of those a few minutes ago), FBI, or some other government agency telling me I’m going to jail if I don’t send money. Fortunately, a huge clue is that these agencies don’t call you on the phone to ask for money. However, I’ve noticed the scammers have stopped using low-quality offshore labor and started using ever-better robotic voice systems that sound far more convincing. (I’m still not sending them any money.)
I can see how the technology has been advancing, and it won’t be long until these criminals will be using artificial intelligence tied to information they have bought from a social network to better take our hard-earned cash from us or change our government.
We have systems that can create voices and mannerisms that can pretty much emulate anyone, with impressive accuracy, and it won’t be long until some criminal figures out how to use them against us.
What we need is an equally capable technology that can figure out what is going on and warn us that what we are hearing is false, and that we should get the hell off the phone.
IBM Has the Answer
Project Debater from IBM is a showcase of exactly what is needed, but at scale. It can listen to an argument and provide a fact-based counter to it real time. It actually won the debate against a professional debater without being connected to the Internet. In production, it wouldn’t have that limitation. It could be that killer application every company desires to have.
One potential application would allow it to monitor electronically what you hear and say, while providing real time advice on what to do. For instance, let’s say you were having a political argument with your sibling, who made a questionable claim. The service would promptly notify you it was false and explain why. Anyone with this tool at a political event could, in real time, point out that the politician on stage was full of crap.
In theory, you could only be scammed by someone who had a far better tool. Otherwise, those who wanted to scam you would have to go through a ton of trouble getting the facts to fit their pitches. They would have to fool your AI.
With that kind of power, they’d be far better served going after bigger fish than we likely are. The cost — and the fact that their AI could be seized — likely would be a huge impediment to its being used for low-value targets.
This tool would be very helpful in litigation. With litigation, much of the cost typically is tied to one side or the other being unreasonable. I’ve been involved with several lawsuits in which I was on the prevailing side, but my award was a fraction of the total legal fees. Both sides would have saved a ton of money had we both looked at the cases realistically.
With both sides having a litigation-focused AI — and both knowing the other side had one — the path to settlement likely would be far quicker and entail much less waste. A judge who had access to this technology would be far more effective at making a determination much more quickly as well.
I’m kind of surprised we haven’t seen legislation to block the use of AIs in politics and litigation, because such systems likely would force a purge of existing politicians from both parties, and massively reduce the amount of money wasted on attorneys.
Wrapping Up: With a Warning
This sounds like a pretty amazing future, but it doesn’t come without risk. Given the power that these AIs likely will have and our eventual dependence on them, we’d likely be even more screwed than we are now if one were compromised. This is a valid concern, because they’ll effectively have massive influence on what we buy, who we vote for, and pretty much every decision we make.
Were our AI to become corrupted, many of our critical good decisions would change to bad ones, and the world would get worse at computer speed — meaning we might not have time to discover and correct the problem. So, it will be critical not only that are these machines are secured, but also that we have several solutions in market reducing the odds that a critical mass of them can be turned against us.
For now, though, IBM’s Project Debater is perhaps the best coming tool to protect us from the fake news that has been polluting the world around us.