Due diligence is needed to ferret out false claims of nearing or attaining AGI (artificial general intelligence).
getty
In today’s column, I examine the ongoing gambit of whack-a-mole when it comes to the increasingly egregious claims that AI has advanced to becoming artificial general intelligence (AGI). You can fully expect that such claims are going to get bolder and more audacious. Trying to find and knock down the falsehoods is a proverbial cat-and-mouse endeavor. The tiniest wisp of AGI fragrance is enough to bolster false sightings and garner outsized headlines by those desperately seeking sensationalism and the accompanying misguided accolades and attention.
Let’s talk about it.
This analysis of an innovative AI breakthrough is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).
Heading Toward AGI And ASI
First, some fundamentals are required to set the stage for this weighty discussion.
There is a great deal of research going on to further advance AI. The general goal is to either reach artificial general intelligence (AGI) or maybe even the outstretched possibility of achieving artificial superintelligence (ASI).
AGI is AI that is considered on par with human intellect and can seemingly match our intelligence. ASI is AI that has gone beyond human intellect and would be superior in many if not all feasible ways. The idea is that ASI would be able to run circles around humans by outthinking us at every turn. For more details on the nature of conventional AI versus AGI and ASI, see my analysis at the link here.
We have not yet attained AGI.
In fact, it is unknown as to whether we will reach AGI, or that maybe AGI will be achievable in decades or perhaps centuries from now. The AGI attainment dates that are floating around are wildly varying and wildly unsubstantiated by any credible evidence or ironclad logic. ASI is even more beyond the pale when it comes to where we are currently with conventional AI.
AGI Claims Are Arising
Everybody is eagerly awaiting the attainment of AGI.
To appease that vast appetite, there are hordes of ardent claimants more than willing to proclaim that AGI has been achieved. You can expect these false sightings to ramp up this year and next year as advances in AI continue to proceed. Though we won’t have reached AGI, the mere suggestion or foreshadowing of nearing AGI will be sufficient to get the chickens clucking and cackling.
Who will be making these false claims?
Lots of suspicious characters, along with some that seem to genuinely believe they are spouting the honest truth. It will be challenging to separate the wheat from the chaff. Plus, the two types of sightings tend to feed off each other.
For example, imagine that a genuine AI luminary says AGI has been achieved (but they are mistaken). This anointed claim will then abundantly spark the fakers to jump quickly onto that heady bandwagon. They can easily escape accountability by merely pointing a finger at the genuine proclaimer, even though the proclaimer is sadly mistaken.
The other direction can handily occur too. Suppose a faker says AGI has been reached. The media might be cautious in going along with the unsupported claim. Assume that many AI luminaries mindfully say that the matter needs to be further investigated before they can render a verdict.
But that’s not of use to the rapid media cycle of real-time responses. So, the media finds one AI luminary that is willing to put a stamp of approval on the finding of AGI. Voila, the media relishes this pronouncement and feels they can freely tout AGI has been reached due to securing an admission from at least one AI luminary.
The bottom line is to get yourself prepared for a ping-pong match of AGI claims followed shortly by a debunking of the claim. Your head and mind are going to suffer from AGI sightings whiplash.
Moving The Goalposts Is Spicy
How could anyone reasonably claim that AGI has been attained even if it hasn’t been?
The cleverest trick is to simply redefine the meaning of AGI. It is akin to moving the goalposts at a sporting event. You move the goalposts closer so that however far you can kick the ball, it goes over them. The same applies to the definition of AGI. An AI maker or almost anyone can proclaim that AGI means something a lot less impressive than the fuller and spirited meaning of AGI.
A prime example of this goalpost shenanigans includes how Sam Altman has opted to move the cheese, see my discussion at the link here. He has repeatedly opted to reword the definition of AGI, doing so to apparently make it easier for AI makers such as OpenAI and others to proclaim they are near to reaching AGI.
The trickery works this way. Rather than indicating that AGI is AI that is on par with all human intellect, you can give yourself a lot more breathing room by narrowing the scope. One angle is to contend that AGI is reached if the AI is on par with a certain stated limited set of intellectual domains. Or contend that AGI is only when the AI is economically potent. Many sneaky mind games of semantics can be used to handily make achieving AGI a lot more feasible.
Keep your eyes and ears open and make sure to unravel what someone intends by the moniker of AGI when they start declaring that AGI has been attained.
Independent AGI Watchdog
A thought-provoking proposal is that we should consider establishing a global entity that would serve as an independent AGI watchdog. The entity would be scanning the globe for claims of AGI having been achieved. After reviewing those claims, the watchdog group would produce an official and reliable announcement about the veracity of each one.
Some believe that the United Nations ought to take on this role. The UN has already been deeply involved in gauging and monitoring the progress of AI, see my coverage at the link here, ergo it seems a natural next step to rely on the UN to sift through AGI claims.
Others argue that a completely independent entity should be formulated. It would somehow not be reliant on any particular country or nation as to its deliberations about AGI. Perhaps renowned AI experts from throughout the world would be consulted, regardless of their national affiliations.
Skeptics are doubtful that any of these contrivances or contortions are needed. Their viewpoint is that we can all readily see whether AGI has been reached. We can do this with our own eyes. We do not need any kind of special entity or group to tell us when AGI is here. The presence of AGI will be plainly evident.
Forget about forming bureaucracies and just rely on everyday common sense to determine whether AGI has arisen.
False AGI Claims Are Disturbing
When I give talks about the latest in AI, including the efforts toward achieving AGI, I almost always get a question that asks why it makes any difference that some might falsely claim having attained AGI. The viewpoint seems to be that if someone makes a claim about AGI and they are shown to be incorrect; no harm has been done. It is water off a duck’s back. We all move on, and the world continues to turn.
Sorry to say, but there are strident reasons to be disturbed over falsehoods about reaching AGI.
A momentous issue is that those claims can be a form of societal and epistemological hazard. Some segments of the population might react rashly to the announcement of AGI. You see, people exist that firmly believe AGI will be the end of humanity. AGI is going to wipe us all out. Those holding such beliefs could readily be triggered by the AGI claim and take actions harmful to themselves and others.
All manners of premature responses could be activated. Envision widespread fear enveloping masses of people. Another extreme could be complacency. Some might believe that with the advent of AGI, humans will no longer need to work and can live a life of great leisure. The reactions are going to be widespread and all over the map.
Yet another concern is the burnout due to false alarms and the proverbial act of crying wolf. You know how that goes. In this context, here’s the deal. A false claim of AGI goes around the world. The story in due course gets debunked. A new claim starts to fly across the globe. It gets debunked. Eventually, people wear out on the false AGI sightings. If we then do really attain AGI, people might be caught utterly off guard and have a worse reaction due to the prior false sightings.
Seeking The Truth Is Perfectly Valid
All in all, trying to reduce false AGI claims has immense value. Preparing for the whack-a-mole phenomenon is critically worthwhile. We must strive to ensure that the truth about AGI is made known and that untruths need to be exposed and smashed.
A final thought for now on this weighty topic.
Albert Einstein famously noted the importance of ferreting out the truth: “As for the search for truth, I know from my own painful searching, with its many blind alleys, how hard it is to take a reliable step, be it ever so small, towards the understanding of that which is truly significant.”
The truth about AGI is a grandly meritorious mission and that’s the absolute truth.